Test Leadership Congress
June 27, 2019, New York, USA
Test Leadership Congress
Video
Aprajita Mathur - Compliance and Agility - How It Can Be Done
Available
In cart
Free
Free
Free
Free
Free
Free
Add to favorites
71
I like 0
I dislike 0
Available
In cart
Free
Free
Free
Free
Free
Free
  • Description
  • Transcript
  • Discussion

About speaker

Aprajita Mathur
Bioinformatics Software Test Manager at Guardant Health

AprajitaMathur is a Bioinformatics Software Test Manager with an engineering backgroundin Bioinformatics. She has more than 10 years of software testing experience.She has worked as a manual and automation tester on a wide range of productslike instrument software, LIMS, web-applications and integrated solutions whichare used in the fields of agriculture, transplant genomics, forensics andoncology. Currently she leads the Bioinformatics Software Test team at GuardantHealth, which builds the world’s leading comprehensive liquid biopsy. Thisnon-invasive tool for accessing and sequencing tumor DNA is used by oncologiststo help advanced cancer patients. She has recently started a chapter for “WomenWho Test” in the bay area and hopes to grow this community. In her spare time,she loves to spend time with her family and write new songs for her son.

View the profile

About the talk

Delivering a compliant product is a resource intensive and challenging activity for most teams. Whether a team is trying to adhere to company, industry, or international standards, it needs to produce deliverables under tight deadlines with the right level of quality.

When you work with Forensic teams the stakes are high. Next Generation Sequencing (NGS) is a new forensic DNA sequencing technology which can result in increased detection ability for degraded and complex mixture samples. It can also provide ancestry and physical trait information which help's narrow down suspects.

Join Aprajita Mathur as she shares how her team successfully built the first Forensics, NGS “sample-to-answer” platform, working in a cross-functional team, using a scrum-based methodology, yet in a compliant environment. For her team, the stakes were high, timelines tight, and quality was of upmost importance to ensure the truth is always found: An integrated “sample to answer” solution to aid forensic investigation teams must be accurate, precise, reliable, and provide information in the timescales of investigation.

Share

Or if something doesn't make sense or you get curious. 00:02 Please feel free to stop me. 00:06 Let me know if you have questions that go beyond the time limits. 00:08 We can always converge during the next 2 days that I'm here. 00:14 So a little about myself. 00:17 I live in California. 00:20 I am very adventurous persons every year. 00:20 I tried to do one activity that I don't want to do. 00:25 And so that's the. 00:29

The hardest one, 00:29 I've done outside of child labor and I also run a woman who test group in San Francisco. 00:33 If you have friends in that area, 00:39 let them know get them in contact. 00:42 And I have 2 boys. 00:44 When is my son and one is my husband? 00:46 I live with them and they are awesome you can also? 00:48 I'll get in touch with me on LinkedIn. 00:52 So what do I do for profession I work for a company called garden? 00:55 And we are in the space of cancer detection. 01:00

We are located in San Francisco. 01:02 Is there a clicker? 01:05 OK, thank you. 01:09 I'm very public company were boring really fast or do we really do. 01:09 So our primary product is a liquid biopsy test? 01:14 What does that really mean? 01:17 Well generally when a patient goes to a doctor in college is to figure out if they have cancer. 01:20 Uh generally the treatment procedures to get do a surgery get a physical sample of that issue from the organ that they think has tumor. 01:26

Our company has developed a technology called liquid biopsy, 01:34 which basically uses blood so you would go into a doctor's office give 2 tubes of blood like you were good. 01:38 Or like cholesterol test or your lipid test right and we use a CF DNA or cell free DNA in your blood. 01:44 To detect very specific areas in your genome and be able to say whether you have a type of cancer or not. 01:52

And In addition to that we are also able to tell which therapies or medicines in the market are best. 02:00 Do that for your Type of cancer, 02:09 so our flagship product is garden, 02:12 360, it's in late stage cancer and we are. 02:15 Working actively on early stage cancer. 02:18 So. 02:22 Where are we going to talk about in the next 45 minutes? 02:22 I tried very hard to put this together in 45 minutes. 02:26 I don't think I can cover everything. 02:30

So again feel free to stop and talk to me afterwards, 02:32 but you cover a couple of things first of all. 02:35 I like to give you some landscape of the project that I'm using as a use case for this presentation. 02:38 Because it's very important to understand what was being built and what were the complications and challenges just because? 02:45 Every product is different even if the same companies building it. 02:52

And then we will talk about agile methodology's that we used. 02:56 Some Agile Bots I call him, 03:00 agile, but because it's agile, 03:02 but and then you talk about things that we learned were not working. 03:04 How we improved it how we improve them again? 03:09 And we thought we were done, 03:12 but we weren't we had some gotchas and then finally we will talk about some learnings. 03:14 Through the process. 03:19 So who here has seen a crime show. 03:19 Raise of hands. 03:24

Really, you guys are watching television. 03:24 No crime show. 03:28 OK so all of us have right. 03:32 And so the project that I would be talking about is. 03:34 Of around crime. 03:38 So my last company that I worked at was ilumina and it was building. 03:38 The 1st next generation sequencing. 03:44 Forensics product what does that really mean but we already have we already fighting crime refining people? 03:46 What is this new solution? 03:51 Why do we really need it? 03:53 So I joined the team. 03:55 And. 03:58

My dreams were shattered. 04:01 So, 04:01 in forensics industry and what we see in Television is very different if you guys saw crime show. 04:04 And there was a crime where a murder was committed there was a lot of blood lot of evidence. 04:10 There you know, people went in, 04:15 they took the samples they came to the lab. 04:17 How long did it take them to bring results for those samples? 04:19 Give me an average number. 04:23 What do you think? 04:25

5 days one day 2 hours. 04:27 52 seconds 04:31 2 days OK. 04:34 Any guesses. 04:34 No OK. 04:38 Generally, in these crime shows right they would go and figure out will bring everything back will put it in the machine and they will get an answer. 04:38 And then in 3 days, 04:49 they will file a court case is done. 04:50 That is not the reality that is not how it happened. 04:53 Even in the best of situations, 04:56 so these are the assumptions that are true to forensics in general so. 04:58

People think that there's enough evidence if you find DNA you will catch the guy not true are girl. 05:03 Sorry. 05:08 Doesn't matter it's not true? 05:08 And. 05:11 Other thing is that the labs US or international they're very high tech like they're using the best of Technology. 05:11 They have all facilities you're fighting it out. 05:21 No, you're not bad man doesn't happen. 05:25 So. 05:28 They don't have super size, 05:28 so this is from the show that I love called bones. 05:30 I watch it a lot. 05:33

I still watch it. 05:33 But they have this they have this cool. 05:33 3D animation thing where they'll get us like a DNA sample and they'll create this. 05:36 Will profile of the victim? 05:41 No. 05:44 No where even close, 05:44 maybe in 50 years from now that is not the reality. 05:46 So what is really true in the world of forensics and what were we trying to build? 05:49 I'm going to go a little biology on you guys, 05:53 I'm going to try it covered really quickly. 05:56

So all of us as humans have 22 chromosomes you get one from your mom and you get one from your dad. 05:58 And you have your sex chromosomes if you had a female. 06:03 You have an ex ex if you are male, 06:06 you have an XY. 06:08 In the field of forensics what they do is you see those. 06:08 Yellow markers basically user below side points and these are very specific positions in your DNA that the forensic scientists used. 06:13 Identify somebody. 06:21 How do they identify? 06:21

Is it a lot of data but? 06:26 So these are your 13 points. 06:30 And each point has a very specific type of sequence. 06:32 So this is from Mom. 06:36 This is from that? 06:37 What does the number 15 mean? 06:37 What it really means is that? 06:40 In one of my chromosomes at that position, 06:42 I had a sequence that got repeated 15 times. 06:45 That's what the 15. 06:48 And the other chromosome. 06:48 It got repeated 17 times and so you get a profile for everybody, 06:51 I have a distinct. 06:56

Profile from you from you from my cousin my dad my mom whoever. 06:56 So, in this case in like testing world. 07:02 This is a happy use case you have an evidence you have a 2. 07:05 Suspect you can see these guys are pretty much exactly the same. 07:09 He saw. 07:13 Awesome. 07:13 That's what would happen. 07:13 In reality, this is what the data looks like. 07:17 What happens is you don't have enough data right so you don't get that 13 points to even say this is the person you only get. 07:20 4. 07:27

Now that could be pretty common. 07:27 Or you might have a weird situation where you have. 07:30 A mixed samples if you have a mic samples. 07:33 Where do you do with that you have too much information you can't really? 07:35 Abstract those 2 things out and say, 07:39 This is from this person or this is for this. 07:41 An even though all of these things exist. 07:44 If I was to get all 13 points. 07:48

With all 2 markers, it's one in a trillion chance that it would be the same with another person. 07:51 So it's pretty robust, but you're still not there, 07:58 so you will solve good cases. 08:01 If you get all the data but not always. 08:02 And then there are some edge cases like identical Twins, 08:05 so for identical Twins, it's exactly the same. 08:08 You cannot distinguish between 2 identical Twins. 08:11 Twin committed a crime. 08:13 And the other twin wants to help him. 08:13

Or it doesn't want to help him because he did it like no, 08:18 he did it. 08:21 And they cannot connect. 08:21 And this is true you can Google, 08:23 it up. 08:25 And the next one that I want to point out is Golden State killer so this, 08:25 dude was in California committing crimes. 08:29 I don't know how long? 08:33 He got convicted after 30 years. 08:34 30 years, he's killed so many people. 08:37 Why? 08:40 What was the big deal, 08:40 he is he's committing crimes all over you should have enough evidence? 08:43

The reason was that he didn't leave enough evidence. 08:46 He was in the IT was in. 08:50 I think it was a police officer. 08:52 And how they found him was through a 4th generation descendant. 08:55 And this was outside of the regular database in the regular points that you use on the on the DNA? 09:01 Whose data that's not generally used does this all makes sense. 09:08 OK So what were We Building? 09:11

We were building the first solution ever in the field of forensic this does not exist. 09:14 This is like. 09:18 IPhone before got launched right in regulatory space. 09:18 We are building a solution. 09:23 You can think about it like a deconstructor. 09:25 So imagine a machine somewhere in the future where you put like a burger in it. 09:28 Turns out, it gives you the entire recipe right every single thing. 09:34 So for our solution, we didn't have just those 13 points. 09:37

We had I remember correctly vowed. 09:41 30 something. 09:45 And In addition to those 30 something we also had what are called snips which are like single small little changes in your DNA? 09:45 So, in total, 200 points, 09:54 so you have lot more data to to play with so if you have. 09:56 Lost evidence or if you have mixtures. 10:01 It's much easier to analyze this information. 10:03 So here's a good example. 10:06

So imagine this in the identical twin queso in Old Technology Cavaleri Electro forces. 10:08 You only see 11 and 12 right for. 10:15 Of this specific position. 10:19 With our technology you could actually go and say, 10:19 Hey, what's inside that 12 what is the sequence inside the 12? 10:25 And you'll notice here. 10:29 There's a see here and the ear see that difference. 10:29 That differentiates between those identical Twins. 10:34 Does that make sense? 10:38 OK, 10:38 cool. 10:40

So you're building this product. 10:40 Nobody knows how it's going to go. 10:43 It's in research mode. 10:45 Ann. 10:45 We decide we want to use Azure. 10:48 So this was more of our team decisions was not something that was forged or we were trying to change. 10:51 This is something a little different than. 10:55 What most companies are trying to it? 10:58 Was it was a joint decision you want to do agile? 10:59

The main reason why we wanted to do agile was because of the unknown because agile allows for unknown, 11:03 we were building a product. 11:09 But we did not know how the customer would react to we did not know what we needed to build. 11:11 Well, we know what we needed to build but we didn't know how it should look how it should interact with the customers. 11:16 We didn't have all that stuff flushed out. 11:22 It was a brand new product. 11:24 It was a brand new market. 11:25

For this company to go into so the main reason we decided was the unknown there are many reasons why. 11:27 Somebody would want to do agile. 11:33 But this is important, the reason I point this out as it's important for you to know. 11:35 Why you need to go after I've had conversations with people and sometimes people say? 11:40 Isn't that what everybody's doing? 11:46

And my answer is maybe but is this the right is this really the right platform for you to use like one is the reason you want to do it. 11:49 So something for you guys to think about. 11:56 So fine we went agile and we are in regulated space, 11:59 the company. 12:03 We were doing this at Alumina. 12:03 It's a leading provider in sequencing instruments that they've built many medical devices before and so for us, 12:06 the regulatory body here was the FDA. 12:13

And that's not a big deal shortly going through this whole process. 12:15 However, for this specific project everyday is not the regulatory bodies, 12:18 so completely different compliance completely different regulations. 12:23 It's a little different an we have our processes. 12:27 We have our procedures. 12:30 We have all of these things. 12:30 All of these things set up. 12:34 So as a team we decided to go agile. 12:36 By my SLP my work instructions may SDLC Lifecycle. 12:38

All that stuff talks about waterfall. 12:42 What what did I miss NBC if you're gonna go agile? 12:46 Sure, my team said. 12:49 You're going to go agile did the companies they were going to go agile. 12:49 No. 12:55 Big problem. 12:55 Now everybody on the team, 12:55 so this is a big team because of all the small components that you got there's. 13:00 A team like chemistry team that's involved in extracting making sure that we were able to extract as much DNS we can. 13:04

There's a team that's building the hardware. 13:12 The sequence or there's a team that's building the software there's uh. 13:14 Team that's building the data analysis or buyer informatics part of it. 13:17 There's lots of stuff happening. 13:21 All these things all of these themes. 13:22 Pink they really think and they believe. 13:26 That they're doing agile development. 13:29 They also think there will doing agile development is the right way of doing agile. 13:31

And to Top of that they think their priorities the highest. 13:38 So a good example would be software is a common thing that's going to be used across right and so. 13:43 The chemistry person comes Hey, 13:49 you know what I really need to start seeing these sequences in the software. 13:51 The bank dramatics teams come and says you know what we started processing 96 samples now have you done any performance testing? 13:55 I want to give you some more data? 14:01

Another thing I want to integrate with the hardware. 14:03 Everybody has their own priority in software team has their own priorities are for things like. 14:06 No, we want to work on the login page where you're talking about. 14:11 So everybody has their own priorities. 14:14 Everybody is doing their own agile, 14:17 and the organization at the high level is. 14:19 Completely somewhere else. 14:21 So that brings me to the elephant. 14:21

There's this huge elephant in the room and nobody is acknowledging it. 14:24 Everyone's like yeah, we're doing, 14:28 it, we're going at all. 14:30 But nobody is acknowledging the elephant. 14:32 So. 14:35 What did we do? 14:35 Be brought it up. 14:38 Multiple times. 14:38 So there's something this is not a one day process. 14:40 This is just something that this team went through we really had to acknowledge the elephant as a team. 14:43 I had to acknowledge that yes, 14:48 there is a problem. 14:49

There is a problem and multi layered system. 14:49 It's All over the place right. 14:53 We had to start accounting for business level artifacts. 14:56 It's a much bigger challenge to change the culture when you are talking at the organizational level right so. 15:00 You're still doing scrum you're still doing agile, 15:08 but there might be deliverables at the business level because you have phase gate exits or whatever. 15:11

We had to start accounting for those because somebody on the team is actually going to spend their time. 15:16 Working on that presentation that they have to talk to senior executives about right so they had to do that. 15:21 We had to start estimating tracking and accounting. 15:26 The most important thing that we did was we assigned one product owner. 15:29 This person is not this person is an expert in forensics. 15:36 They're not an expert in chemistry. 15:39

They are not an expert in software development, 15:42 so if you want to ask them what's the best way of. 15:44 Architecting software, they may not know, 15:46 but they know the vision. 15:49 They know where this is headaches, 15:50 or one single person who decides. 15:52 And have the authority to say yes or no right. 15:54 Be assigned scrummasters so these individual teams got one scrum master eating. 15:58 This could be anybody. 16:02 This doesn't have to be the software development lead or? 16:02

It literally could be anybody who's willing to do that work right. 16:06 And some point of contact. 16:10 So if you have integration for psych testing is one of the things that just needs to happen across the board, 16:12 so there has to be a point of contact. 16:17 But then this was advertised and talked about that. 16:19 Everybody understood that this is the structure irrespective of the structure in your? 16:22 I might be reporting to whoever does not matter. 16:29

When the product owner says is priority, 16:32 one that is priority one. 16:34 That makes sense. 16:36 OK, cool. 16:36 So we move on. 16:39 The next thing that we had to do was before anybody could do any development. 16:41 So this is this is through the process. 16:45 It's not like we haven't started developing a product development already starting your like. 16:47 Starting to do stuff but we're realizing all these things are happening. 16:52

So we had to start alignment at a higher level, 16:56 the individual team owners had to align first they had to. 16:59 A line before they could say Hey. 17:03 This is what you're building so it's a product owner's decision to say OK this is his priority. 17:05 Once you have that. 17:10 You align and these teams are different goals. 17:13 And different Springs did not not necessarily the aligned timewise, 17:16 but every team had us printable this is what we're trying to achieve. 17:20

And there was a lot of negotiations, 17:24 so this is not an easy meaning where somebody hands you down the list of things that you want to go. 17:26 Nope. 17:31 Lot of negotiation during the planning meeting so this was really the product owner understanding the complexities are trying to understand the complexities. 17:31 That these dreams are trying to fight right. 17:41 Ann. 17:45 Of getting on the leads involved so getting different. 17:45

Eyes on this not just the scrum master, 17:49 but somebody from test. 17:52 Somebody from a user design somebody from marketing so that everybody is aligned. 17:52 Once we did that once we did that pre work. 18:00 The team started to crunch. 18:03 We were We were spending before that, 18:05 but they're crunching their spring being were doing. 18:06 Awesome it's beautiful and you're done. 18:09 Presentations over we have a beautiful scrum, 18:13 Agile Model Grace can come home. 18:16 I'm just kidding. 18:19

We think you're doing awesome right. 18:19 You're doing these demos, we have these retrospectives here doing all all things that you can think our activities that you need to do during scrum right so it's. 18:23 It's beautiful, it's working perfectly. 18:32 Everybody is really happy. 18:35 Then one day. 18:35 We have a demo. 18:39 An one night before one of the developers. 18:39 Things me and he's like. 18:43 Hey, a project that you know that feature that I. 18:44

That I developed this one, 18:47 I need to demo that. 18:48 And I'm like OK? 18:50 What do you need then it was like OK, 18:50 you know what I I need this specific type of samples and I could demonstrate it. 18:53 Or I can populate it for you, 18:57 OK, bye goodnight Frank. 18:59 Nice name the demo. 18:59 Uh right before the demo. 19:03 I tell him he did, 19:04 just remember to click on sample. 19:06 A He's like. 19:08 Hey, I got this, 19:08 I got. 19:10 So he went in and he starts doing this demo, 19:10 he clicks on symbol B. 19:13

And the software comes crashing down. 19:16 And everybody in the room is like. 19:18 These are stakeholders their product is really. 19:23 I thought you guys were like 2 weeks of what was happening this chaos. 19:26 Everybody is terrified we somehow could get out of that demo that retrospective was the bet best retrospective that team. 19:31 Ever. 19:39 Skeletons literally. 19:40 Came out of the closet. 19:40

Testing was upset they were like you guys give us a development build 2 days before were supposed to do the demo? 19:44 How are we supposed to test it the developers like? 19:49 This is took too complex. 19:52 I somehow got it. 19:53 Working we have 5P one bugs in the backlog were you guys talking about this is not really working. 19:53 So. 19:59 You are kind of faking it. 19:59 We were making it look good, 20:02 you are making agile good. 20:04 We were making. 20:06 Software development. 20:06 But. 20:09

In a compliance environment. 20:09 Is it really a shippable product? 20:12 No, it isn't. 20:14 So what are we shipping I don't know have you guys ever experienced this where everything is almost done? 20:18 Right we can relate to that. 20:26 OK, perfect. 20:29 That is exactly what we were doing. 20:29 We were getting everything almost done right, 20:32 but we were burning it. 20:34 City was beautiful, it was consistent it was amazing. 20:36 So once he's tried that and once we realized. 20:42

We had to really think about what are we trying to do? 20:46 An that's why I say you have to build it like you mean it the only thing that if you want to take. 20:49 Out of this, this presentation. 20:55 Is to build a shippable product? 20:57 In your regulated environment if it's not if you can ship it in that environment. 21:02 It's not shippable. 21:08 You see what I mean? 21:08 It is not an update to a Facebook page where you can now do a heart instead of a thumbs up right. 21:11

It impacts somebody's life or impact something else right. 21:17 So we started getting the team involved. 21:20 This is probably the hardest thing is cultural change getting the entire team. 21:23 Accountable for that product and this takes time, 21:29 it's not going to happen overnight. 21:32 But we started to add. 21:35 Something called acceptance criteria for a story to even make it into the Sprint. 21:37

So imagine the situation before your Sprint planning before your grooming like after your grooming and before your spring planning. 21:43 If there's a feature login page. 21:51 The team responsible for working on that feature is going to go sit with a scrum master 2 minutes meeting. 21:54 And he's going to pull up the story and they are going to discuss it. 22:00 And they're going to say OK developers wouldn't say I have all the requirements that I need. 22:03

Pastors gonna say. 22:08 I have everything I will wait I don't have the data. 22:08 You cannot pull it into the Sprint. 22:13 If this team says it does not have all the information it needs to get that thing done you do not. 22:16 Pull it into the spring, 22:23 you don't even discuss it during planning. 22:24 It's strong Masters responsibility to go to the product owner and say, 22:26 I'm sorry I messed it up. 22:30 Not the team. 22:31

That was quite crucial the second thing was definition of done. 22:31 And this is very, very important from testing perspective. 22:37 Because we had to start doing definitions or automation, 22:40 I reprint automate this feature. 22:44 Is everybody is everybody in agreement? 22:47 Is the developer going to make it auto Mabel? 22:49 Are they going to put components in the code so that it's easy to automate? 22:53 Or are we just going to manually test it. 22:57 Everybody had to agree on it. 22:59

And. 23:01 Did it up we did a lot of things but in this included documentations if you're if we are? 23:01 Like if this is the final aspect of the feature. 23:07 It has to be documented and done right, 23:10 so that is also part of the definition of. 23:11 But the the. 23:14 The impact that definition of John had on this team was. 23:14 If every single thing. 23:21 On the definition of done is not met. 23:21 You cannot demo it. 23:26 That was a very hard sell. 23:30

So this took us some time, 23:32 but it's really the cultural and the behavioral change within the team members of supporting. 23:34 Each other that got us to the point of being able to ship it to the level that we did. 23:39 So. 23:44 That's all agile little talking compliance. 23:44 Sorry I completely skipped the slide just because I'm very close to my. 23:51 You're developing this we're almost there. 23:56 We have all these features were starting to doing. 23:58

UAT testing not have to go talk to a regulatory team member or quality team member in my team. 24:00 To be able to ship this right internationally or anywhere, 24:05 so that I made all the regulatory standards that. 24:07 That exists. 24:10 And this is what they told me. 24:10 I've never done agile. 24:14 I don't know how to work with this. 24:14 I don't know how to put the submission. 24:18 It's a problem. 24:20 I don't know I don't even know if they supported. 24:20

So this is what I generally point I know this is from FDA and there are people from outside the FDA. 24:24 So this is what it means. 24:30 FDA or these regulatory bodies they understand the value of doing concurrent engineering. 24:31 I'm not going to read that. 24:36 But what what it basically means is building something constantly, 24:38 and getting feedback right that they acknowledge that fax is not and you. 24:42 Concept so how do you do it in the agile model? 24:46 So. 24:49

This is my way of presenting it. 24:49 You need to go back be able to go back to your organization and say, 24:53 Hey in waterfall through design inputs. 24:57 For design how inputs this is how it Maps to agility. 24:59 This is how I'm doing that work. 25:04 I'm still doing that work. 25:06 I'm just not doing this in 4 blocks. 25:07 I'm doing it. 25:09 Constantly, how. 25:09 So when you do product planning Sprint planning Sprint retrospective those are all design and development planning stages. 25:12

Right they're just doing it all the time. 25:19 Design inputs, those are your product backlog. 25:21 The Sprint backlog so user stories all of those are your design inputs for design outputs are. 25:24 Acceptance criteria testing and you are doing cordon design reviews is our design reviews design changes you're getting customer feedback. 25:32 You're demonstrating this to your product owners, 25:40 you might even be doing. 25:43 UAT you might even be doing validation. 25:44

All of that stuff is part of your design change and you're doing risk management because every single Sprint you're discussing the priorities of the features. 25:47 We're doing everything on that list that a regulatory compliant body would ask you to do. 25:56 So we got there, however. 26:04 The biggest chunk for any organization or any team would be to optimize. 26:08 So. 26:13 If you have to optimize optimize behavioral changes just takes time. 26:13

But if you do it right you get your team in that mindset. 26:18 You can do it second is you need to build your process you really need to know how you're going to do it and you need to. 26:22 Align people on that process. 26:29 It might look different from what your company's doing right now. 26:30 You need to know what you're doing you can change it. 26:33 But keep everybody involved. 26:36 Build your tools. 26:36 So we use all sorts of tools. 26:40

We were using Jira Confluence, 26:41 JAMA, you name it. 26:43 But we didn't use them out of the box. 26:43 We had to customize it. 26:46 For whatever process we defined we will use it. 26:47 How to customize this if you have to put some investment into it? 26:50 And build your work was like, 26:54 how are you going to do? 26:55 How are you going to manage so if? 26:57 There's integration testing that needs to happen. 26:59 How are you going to align that between 4 scrum teams you have to plan this out? 27:01

So. 27:07 Why did we start working on this project in the 1st place because he wanted a better product or the customer? 27:07 And. 27:13 Be multiple times it happened that we lost sight, 27:13 so remember the Y remember why you're building something. 27:18 So, in this case. 27:21 Because we wanted to constantly be getting feedback right. 27:21 We were performing UX testing or you UX workflow testing. 27:27 And what that would mean is when we had uh features developed correctly. 27:31

You could put it in a room. 27:35 And we would invite researchers scientists and other collaborators who come in and try our software and submitted somebody would just sit and watch them. 27:37 And that gave us a lot of feedback. 27:45 We shared the software with internal and external customers so it would. 27:47 Sort of validation right so they're performing like a user. 27:50 These are all activities that you can do in parallel without impacting and still receive. 27:53

Feedback and improve. 27:58 Clarify everything clarify for customers if you're not sure why something is happening the way it's happening clarify. 28:02 Acceptance criteria for your stories to make it into a Sprint so this is a hard mark for what your team works on you have to hold a very high 28:10 level. 28:18 And a heavy high standard you also have to have a higher standard for definition of done So what did we learn today? 28:18

You need to define your compliance or company might be following something else. 28:26 You might be doing something else. 28:30 You need to find the Delta and you need to bridge the gap. 28:32 If there is a gap right and you need to bring them on board. 28:34 This is the challenge. 28:37 I'm facing this is how I'm going to solve it. 28:37 Define or agile, you need to work with your team. 28:41 Oh, I completely forgot. 28:44 I forgot to mention something. 28:44

We also got scrum trained as a team together. 28:47 And that's hard because people had their own vision of how they were going to do scrum and I'm not saying one way is better than the other. 28:50 But that just helped us a line as a team of what acceptance criteria means? 28:57 What definition of dance mean? 29:01 What is a product owner what? 29:03 How does a scrum master? 29:04 Work so that training actually really helped. 29:06 So we define how we were going to do agile. 29:08

We were not just making shippable products. 29:11 We were trying to keep customers at heart. 29:13 And perform constant feedback. 29:16 Uh during the building process so we could deliver the best product. 29:16 Get the team involved early so I hear a lot get the tester involved early no get the team involved where you have to think as a team if you 29:23 don't think as a team no matter. 29:31 How many testers you have on the team it's not going to work? 29:33

Account for all deliverables if you're in the regulatory environment and you have to ship with 52 documents. 29:36 Account for that on your backlog. 29:42 Trump cannot fix culture. 29:45 Is this a mechanism that can improve it? 29:49 So if that's what you're trying to achieve is going to be hard. 29:52 And change is hard and it takes time, 29:56 so be patient with your team continue to do what you're doing. 29:59 And ask for help. 30:04 When needed. 30:04

Cackle comments for the website

Buy this talk

Access to the talk “Aprajita Mathur - Compliance and Agility - How It Can Be Done”
Available
In cart
Free
Free
Free
Free
Free
Free

Access to all the recordings of the event

Get access to all videos “Test Leadership Congress”
Available
In cart
Free
Free
Free
Free
Free
Free
Ticket

Interested in topic “Software development”?

You might be interested in videos from this event

September 28, 2018
Moscow
16
156
app store, apps, development, google play, mobile, soft

Similar talks

Lina Zubyte
QA Consultant at ThoughtWorks
Available
In cart
Free
Free
Free
Free
Free
Free
Christine Fisher
Manager, Business Analysis and Quality Assurance at NAIC
Available
In cart
Free
Free
Free
Free
Free
Free
Priyanka Halder
Head of Quality Engineering at Goodrx
Available
In cart
Free
Free
Free
Free
Free
Free

Buy this video

Video

Access to the talk “Aprajita Mathur - Compliance and Agility - How It Can Be Done”
Available
In cart
Free
Free
Free
Free
Free
Free

Conference Cast

With ConferenceCast.tv, you get access to our library of the world's best conference talks.

Conference Cast
541 conferences
21379 speakers
7781 hours of content