DEV Community

Gaurav Singh
Gaurav Singh

Posted on • Originally published at automationhacks.io on

10 years of Software Testing and Automation

Career progress

Source: Illustration created by Katerina Limpitsouni on Undraw

Today, 8th June is a special day for me. It’s the day that I complete 10 years as a Professional Software tester and automation engineer.

This long post is a trip down memory lane where I call out some of my key personal and technical learnings, experiences, challenges and career highlights.

Disclaimer: This is a gonna be a pretty long one, since 10 years is no short time! Grab a drink or popcorn 😉

How I got into Software testing

It was a bright day in June 2011 when I joined Accenture, my first Tech company fresh out of college with an engineering degree in Information Technology. The vivid memories of that day are still quite fresh in my mind as I remember going through the onboarding process as a complete newbie.

I had no clue about which role, team or skill i’ll get to work on in my career. Ever since college I wanted to be a Developer and had built simple Microsoft .NET and Java websites as part of my college projects.

I liked to code and felt a certain charm towards the role, and thought of it as an elite role to have at that point

My employer did a random skill allocation upon joining and I saw my skill was going to be Functional Testing. 😲

With no clue about what kind of career could be made in Software testing, I remember being a bit disappointed at that point and inquired on whether I can change my skill, However, I got to know that I can change the skill only after spending 18 months with the company/project that I get enrolled in 😲

A first taste of Software testing

Accenture had a Green field training (GFT) program where freshers would be trained on their allocated skills and I was trained for a couple of months on HP QC (Quality center), HP QTP (Quick test professional) using VB Script, Load testing using HP LoadRunner and a foundational knowledge of different types of testing

At that point I was surprised to see that "Okay, even Software testers code?" 🤷 There is something called as Automated testing that provided testers a way to automate away bulk of their testing.

I came to learn that testing is not about clicking some buttons around but involves different flavours like Functional, Non Functional, Security, Accessibility, Usability and you can go deeply into any of these topics

I remember saying to myself

“hey, this looks really interesting, Are you telling me that i get to do all that on a project 😆, seems like quite a challenge”.

After completing my training successfully I joined the ARISTOS DSL team within accenture that built solutions for a major US Telecom client.

Learning: 💡 Software testing is not just clicking some buttons (as is commonly misunderstood), but a holistic field involved in checking if the system works at all of its different layers. It involves not just understanding the system deeply but also writing cool automation to make this whole process faster

Hands on Test planning, execution and automation 🙌

I was fortunate to have really good and inclusive mentors who took me under their wing and showed me how the entire process of testing happened

  • Designing a Test strategy
  • How to write detailed test plans with expected scenarios, conditions and results
  • How a scripted test is written covering the application
  • How does tracking entire execution/defect cycle on a test management tool happen

Experiencing the thrill of becoming a bug hunter 🐛

I really started to enjoy the process of investigating the application and the underlying database structure, infrastructure and while initially I followed test plans created by senior engineers, pretty soon I ended up creating Test plans of my own.

Finding a bug and then getting it fixed by talking with Devs gave me a rush like no other , knowing that the product is better off and I’m directly contributing to increased customer satisfaction 👍

I also learned first hand how Developers, Testers and Technical/Business Architects, Design and other functions, work all together 🤝 to deliver quality software.

Those were the days of long running projects and structured change requests (read Waterfall model) and I saw this whole cycle as it happens in enterprises

Automating tests in QTP 👨‍💻

I also realized that while the process of exploratory testing was very creative and interesting it took me a long time to run these cycles manually.

My Test lead and manager wanted to see if this could be automated.

I started to write Test automation using QTP , initially using record and playback feature of the tool to write data driven E2E scenarios on the application. They were flaky as hell since every time the website changed, I had to go back and modify the identifiers in the Object repository , it was painful to do this and I wondered if this could be done in a better way?

We had a senior engineer on another team who was more into VB Scripting and had written a keyword driven framework wherein instead of using record and playback, each component was Hand-coded in VBScript

After learning and trying this for my own project I found this to be very efficient as managing tests was much easier.

Learning 💡: While record playback technologies are good to understand the application and generate scripts quickly, its not maintainable in the long run and its better to take that knowledge and hand craft the automated suite

Services and Load testing

After spending an year writing UI automation, I moved towards using SOAP UI for Web, services testing by hand and again it was fascinating to learn since there was no UI to play with and you can run functional tests much more quite quickly

Learning: 💡 API testing is another way to verify functional and non functional aspects of the system and is much faster and more reliable to conduct than UI testing

I also got a chance to transition to a team that was running load tests on the application using Apache Jmeter and I realized this was a quite different aspect of testing.

I mostly ran these tests and collected metrics while assisting an experienced Performance engineer and helping to communicate with infra teams to ensure these load tests went fine and the system did not break down

Becoming a mentor 🧑‍🏫

Towards the end of my time with Accenture, I had transitioned from a fresher to a more seasoned engineer and mentored multiple engineers who joined the team

I found the whole aspect of teaching someone on the team made myself learn things a lot better and I had lot of fun talking with them, learning from them and enjoying countless team lunches and outings.

First Job change

After 3+ years with Accenture I decided it was time to move on and change jobs and see how other projects and teams were solving similar or new problems.

I had an offer from an early stage startup and a more stable enterprise company and decided to join Aricent (now Altran) as my next employer in September 2014.

Testing an analytics product

I joined Aricent as a Senior Engineer - Testing as a consultant in the Itron Analytics team which was responsible for building analytics dashboards for Gas, Water, Electricity meters for a big energy and utilities company in US.

When I joined the QA team size was 3 including a people manager.

It was exciting since I had previously not worked with an analytics product and I got to understand how complex ETL processes would take data and seed and store aggregated results into Data ware house tables.

After joining this team, the senior engineer moved on I became to most senior IC on this team.

I specifically learned how to design test strategies for this complex product and also participate actively in all Agile ceremonies working with a cross functional team spread across India and US.

Understanding how a tester can effectively help shape product direction by calling out missing requirements, clarifying acceptance criteria and help team improve by bringing up process inefficiencies in Team wide retro was a lot of fun

Building UI automation with Microsoft Coded UI

One of the biggest challenges in this team was when the client management decided to undertake rewrite of old Flash based dashboards to more rich modern graphs using D3 and HTML5.

The test manager on the team also decided for us to use Microsoft CodedUI as the UI automation framework of choice since we were heavily intoMicrosoft stack (Visual studio, C#, TFS (Team foundation server), and MTM (Microsoft test manager)).

Coming from my QTP/UFT background, it was a challenge to understand C# and CodedUI technology however I was able to learn these and become a productive contributor on the team under a more experienced Lead’s guidance.

CodedUI heavily encouraged use of Hand coding while building the framework and I remember making use of Hexawise tool to generate lot of test combinations which formed the basis of our automated regression suite

Learning: 💡 Don’t be bounded with the Tech stack that you are used to, learning a new automation stack helps you to become a more well rounded Automation engineer since you either see new patterns or new ways of using existing patterns

First onsite to USA

I also got a chance to travel to USA to San Francisco, California and Spokane, Washington along with a colleague to help test and deliver this migration feature under a time bound project

Working in a small sized company

After contributing to this teams success, I wanted to move on to a more challenging role and work with a small team/company and switched jobs to Projectplace India which was recently acquired by Planview USA in 2016. When I joined this company, We had a team of 3 Automation engineers in India that grew to 8 people in some months time.

This role required me to learn Python 🐍 for test automation and looking back now joining this company was one of the best decisions I took because:

  • My colleagues were all seasoned automation engineers with 10+ years of experience and I learned so much from their existing knowledge
  • I was challenged on an individual level to build framework components and utilities daily

I really learned a lot from this experience

Building API automation frameworks with scratch

We liked to call ourselves the Jedi team, (a name i proposed driven from my fondness for Star Wars trilogy) and we augmented a largely manual QE team in USA by building automation tools and frameworks

Pairwise testing

I got a chance to contribute to building a framework that used Microsoft PICT tool for generating pairwise test combinations to test SSRS (SQL Server Reporting Services) analytics reports and performed XML comparisons against a set baseline.

This solution also stored these test results in SQL Server database and had an in house dashboard to present these test results using Jenkins for CI.

I got to see and learn how to build a rich framework from scratch and contributing hands on was a really good learning experience

I still remember, when the India head came over to my desk one early morning and called out the importance of treating oneself as a Software craftsmen and taking pride in being a solid engineer and building these reusable solutions to make QE more efficient.

He was really inspiring and this experience did shift my mindset a bit more to someone who cared deeply about the quality of his code. We frequently paired and did lots of team code reviews to ensure our code was top notch

Coming to office every day 🌞 and coding till evening 🌔 while building automation was very addictive and I believe everyone should go through an experience like this.

Learning: 💡 Working in small teams and companies can really challenge you and give you an opportunity to learn and grow way faster than in enterprises

We also built a custom test framework to test SOAP, REST based APIs and check integrations between different Planview and Projectplace products as well as external tools like JIRA, Rally etc.

Second travel to USA

I also got the chance to travel to Austin, Texas USA for a couple of weeks to understand more about the product as well as conduct trainings for the QE team there on using and contributing to the API automation solution that we had built in bangalore.

Building Web UI automation framework from scratch

After spending a considerable amount of time building API frameworks, I took ownership of doing a POC to build a Selenium Python based framework for web automation. It was a challenge since I had never worked with Selenium before and I read the code from another team and tailored their framework for Planview team.

Learning: 💡: Reading others code is a quick way to improve your craft as a Software engineer, the insights that you gather are invaluable

The learning experience was immense. I got to implement Page object pattern and learn Pytest as the test runner of choice while also setting up CI on Jenkins and a local Selenium Grid setup on docker. It took me around 2 weeks to setup this POC

Taking on this E2E undertaking while learning how these technologies worked really gave me confidence in my skills as an Automation engineer, I then mentored and on-boarded couple of other engineers onto this framework

Learning 💡: Using/contributing to existing frameworks is great but you should really challenge yourself to understand how the different framework pieces work together and build stuff. The more you do this, the better you get with passing time

Taking notes

I saw one of my senior colleagues had this practice of taking notes, whether it be tidbits to solve a problem, snippets or even more varied notes. He was able to access these notes on demand easily and this inspired me to maintain a healthy note taking habit myself. I experimented with different tools like Evernote, MS OneNote, Notion before finally landing on Plain text note taking in Markdown files. You can read more about mynote taking system here

Learning: 💡: Being able to take detailed notes and maintaining a second brain is very important for an engineer as it ensures that no solution approach is lost. Our brains are meant for creative thinking and not really information storage and retrieval.

Attending my first ever Automation conference

I remember attending my first ever Testing conference, Selenium Conference, Bangalore in 2016 and was awestruck seeing and hearing amazing and knowledgeable speakers including Simon Stewart, Dave Haeffner among many others.

It was a great learning experience and I wondered how it would feel like to know enough about something to deliver a talk in front of so many people

Learning: 💡: Try and attend tech conferences, you often would learn about lot of new tips, approaches, tools and techniques, thought patterns about things that you either know currently or that you do not know at all

Working in a Start up product company

It was 2018 now and I saw one of my college friend post messages on facebook promoting his current employer Gojek

While things at my current employer were excellent, I was a valued member of the team and still had lots to learn, I wondered how would it be like to work in a Hyper scale product company that created so much social impact in the lives of South east asian people

I got a referral and after interviewing with them, realized that Gojek could be my next home.

Gojek had native mobile apps on Android and iOS and primary a JVM based automation stack using Appium, Cucumber, TestNG and RestAssured.

I saw this as a opportunity to learn and grow myself more by understanding the challenges with mobile automation and well as getting to learn a popular static language like Java and joined Gojek in Feb 2018

The challenges of working in a small team

When I joined gojek, I could instantly sense a change in pace, we had Bi-Weekly mobile releases that required regression runs on consumer facing apps to ensure nothing broke for our customers and my team had only a single QA consultant from consulting company and one engineer in Jakarta.

There was zero app automation and the API automation done by another engineer (who had left the team) was not managed and broken.

At that time, We also had a very less QA’s as Full time employee and I was the 3rd QA in the bangalore office. 🤯, we did have a large QA consultant team who helped in taking care of quality and automation across different teams though

The GoSend team was also going through the process of redesigning the app based on latest design language. I remember thinking to myself, “Wow, there is so much that I can do here.”

I started building context about the team, the apps, backend architecture as well as existing automation frameworks

Building Mobile Android UI automation

I first decided to improve upon the bottleneck of having to do slow manual testing cycles on mobile apps for the Bi-Weekly FCT (Full cycle test) and started exploring the existing optimus framework which was built by a consulting company called Test Vagrant.

This was a good experience since I got to read code from other seasoned automation engineers and since I was grokking Java at the time, having this structure already in place really helped

Learning this framework architecture, Appium API, Cucumber was an interesting challenge, and within a month or two I had a test suite in place to run against the Consumer Android app on mac mini’s with Gitlab CI runners

I did this automation in parallel to manual exploratory testing to ensure that the app redesign release went fine and this suite reduced a lot of the manual testing effort that I had to put in for this regression cycle

Learning: 💡: If it’s boring and repetitive, automate it 😉

Building API automation using Kotlin, TestNG

After an year of working on mobile automation, I had another colleague join from a consulting company and I on-boarded him on the mobile framework and wanted to set out and build the API automation framework for the team

We had an existing test suite that covered few E2E tests however there were multiple problems with this suite:

  • Tests were heavily coupled together
  • It was very difficult to follow the code or modify to add new features/tests.
  • Tests often would fail and made debugging very difficult.

After analyzing the codebase I decided it was a lost cause to try to refactor this in shape and not to give into Sunk cost fallacy, A rewrite was in order 📈

Choosing an approach

We had an existing framework in Gojek over rest assured that many different teams were using, and while initially I thought to write my own python based test suite, I realized after a week of POC effort that integrating my tests/utilities for other teams would be very difficult since they were mostly using Java based stack and I would have to write lot of bridge code using Jython or implement lot of APIs.

Learning Kotlin

At this point, there were few developers who also wanted to contribute to the framework and automated test suite and suggested to try usingKotlin as an alternative to Java.

Kotlin at that point was already the most popular language for Android development and had already been blessed by Google as so, After researching and learning the language I found, that not only writing code in Kotlin was faster, but it also was a lot more concise and expressive than Java. I wrote couple of blogs on Kotlin features. You can readPart 1 orPart 2 on my blog

I decided to use it as the language for my team and convinced my teammates of the same. It learned it from Kotlin Koans and other courses on Udacity and more so on the job.

Also, wrote a lot of blogs on using Kotlin along with TestNG. You can read themhere

I later extended this test suite to run for all geographies that Gojek was in as part of an App unification project and reduced the suite execution runtime from 1.5 hrs or so to 25 mins by implementing concurrency via TestNG and Kotlin synchronization blocks.

From Lone QA to becoming a lead and then an Engineering manager

For the first two years of my career at Gojek, I was mostly a Senior IC (Individual contributor), designing test plans and automation myself, as well as directly mentoring one colleague.

Lone QA

In fact for some 6-8 months I was the only automation engineer on the team.

How did I manage to accomplish what i did with such people crunch?

Well, I mainly ensured testing happened in the team regardless of who did it. On our team developers tested lot of their own features and were very supportive of me. It involved understanding them and building a good relationship based on trust and mutual respect.

I would usually pair with Devs on the most critical features, helping to brainstorm possible coverage gaps while planning and building automation solutions and this simple model led us to deliver many features to our customers with high quality

Learning: 💡: Learn to trust people around you (in general) and more so if you are the onlyQA/Automation engineer on the team, You won’t be able to gate keep every change, but can surely sensibly automate flows that add maximum value and give dev’s the insights and tools to do more effective testing.

Becoming a lead

Come 2020, With Covid pandemic and work from home, our team had to ramp up to adjust to growing needs of the logistics business and the we decided to hire 2 more automation engineers

Managing the hiring and then onboarding process was a tough process but I along with another senior colleague were able to achieve this by setting up well defined process, setting proper expectations as well depending heavily on documenting the system.

We did lots of pair programming and code review sessions to help the new joiners become comfortable with the team

Being the senior person on the team, I was also asked to manage these colleagues directly, This was a shift in my approach and mentality since while I had lot of experience mentoring and working with team members, I had never really done direct people management before.

I mostly managed this by setting shared vision, setting up processes and offering an empathetic ear during 1:1’s

Becoming a manager

Within 8 months of me directly managing a couple of engineers, I got promoted to being the Quality engineering manager of the entire logistics product group. I must admit I had my doubts about whether I should continue to pursue the path of Staff engineer or try engineering management and though the Jury is still out on what I would really like

I feel that learning to be an Engineering manager could add more dimensions to my personality as well as skills.

The key difference that I feel is that now I can take a more active role in shaping up the quality of the product by using multiple tools at my disposal all the while helping my colleagues excel in their interest areas.

Stay tuned for more thoughts on this topic in the coming months

Becoming a blogger 📝, Conference speaker 🗣️ and Technical course instructor 👨‍🏫

Setting up a personal tech blog

Around May 2018, I picked up another habit of listening to multiple podcasts on the bike ride to office and I started withThe stack overflow podcast with Jeff Atwood and Joel Spolsky

I really enjoyed their conversations around how Jeff was building stack overflow and the challenges in scaling it.

I also started following and reading their blogs and felt an immediate connection with Jeff’s writing. There were many engineers in Gojek who had their own personal blogs as well as blogged heavily on the company blog site

I thought to myself:

Blogging seems really interesting, maybe I should give it a shot.

There were many inhibitions that I had in my mind, whether anyone would read it?, what kind of blogs would i write? Jeff calls some of these fears out inFear of writing article

Instead of overthinking it, I just started with it. Setting up a blog on Wordpress was really not that tedious, I decided on automationhacks as a name, since at that time I was reading some blogs on life hacks and found that tidbits and insights into solving problems really saved me a lot of time.

Also, I wanted this blog to capture a breadcrumb trail of all my learnings over the years while solving different problems. You can read my blog automationhacks.io

Learning: 💡: Blogging about your learnings is a really cool way of firstly clarify if you yourself understand it, plus other people usually benefit a lot by hearing your perspective about solving the problem. If you are in doubt, don’t overthink, Just start 🏃

My first conference talk at Appium Conf

It was around Jan 2019, and I saw the CFP (Call for papers) for Appium conf open up. As I mentioned before I had this goal of delivering a technical talk at a conference. I wrote up and sent my CFP not knowing whether it would be selected, but still by now I had realized its better to try and fail rather than not try at all.

To my surprise, my CFP was selected and I delivered my first ever talk in-person in Appium Conf, Bengaluru. If interested, You can read about this process of writing a CFP onGojek tech blog.

I had a great experience at the conference and met world renowned engineers like Angie Jones, Jonathan lipps, Wim Selles, Jason Huggins in person. These people are regarded as legends in Test automation space and I was a speaker along with them. What a feeling! 🙌🏼. I wrote a blog with my Appium Conf experience, If interested you can read more about it herehere

My first tutorial course: Visual validation using Applitools

I had heard about Test automation university by Applitools and taken a few courses

During the Conference I got a chance to talk to Angie Jones in person during AppiumConf, I asked how can I help with this awesome initiative?

Based on my python background, Angie mentioned that the platform needed a course on AppliTools Python SDK. To be honest, at that time I had not tried visual testing hands on but I signed up, knowing fully well that I could learn and figure it out. 😉

Learning: 💡: When someone says, can you do something? (If its interesting to you) Say yes! And then get busy figuring out how you can do that. In other words. Fake it till you make it! Trust me, You’ll learn a lot more than saying Yes that waiting for the perfect moment where you have all the details figured out

I followed Angie’s tutorial on Applitools with Java and extrapolated it to Python language by reading the Python SDK docs and created the course.

Also, Recorded these in the Gojek Bangalore office in early office hours (my home was in a very noisy place) over simple Sennheiser headphones with mic and edited it using QuickTime and iMovie as editing software. I got useful feedback from Angie during the production process and I learned so much about producing video courses.

After finishing this course and seeing my name live on the siteamong many other world renowned instructors.

I could not have felt more proud of myself. 💁

Learning: 💡: Most of the things that you want in life, could be achieved when you put yourself into uncomfortable situations and by talking and working with other people! Don’t be shy. Go and put yourself out there.

Speaking at Online Automation Guild conference on Contract testing

Coming fresh from my first talk, I wanted to try my hands at different conferences and I saw CFP for Automation Guild being run by Joe Colantonio, I had been a big fan of Joe’s podcast thenTest talks , (now called Test guild automation podcast ) and decided to see if I can speak at this prestigious conference.

I remember hearing from someone in aministry of testing podcastthat one of the best ways to learn something new is to give a talk on it.

I had heard about Consumer Driven Contract testing but really wanted to learn it and see if thats something I could use in my own team and company

Proposed this as my topic and again to my surprise was accepted as a speaker (I guess, somehow I was able to write good CFP’s 🤷).

I prepared and recorded this talk while learning and building knowledge on CDC using PACT in WeWork offices (which gojek was using as a co-working space) on weekends and my recent experience with building a course really helped since I was already familiar with the process.

During the actual talk and post talk Slido questions, I got lot of interesting questions and they also helped me understand what aspect of CDC I needed to learn more

Learning: 💡: One of the best ways of learning something is to teach it to someone else

More talks and courses

I later on went to give another talk during Selenium Conf Online 2020 on“How to build an automation framework with Selenium: Patterns and practices”and built another course onAPI testing in python for Test automation university. You can see entire catalogue on my bloghere

Any regrets in these 10 years?

So reflecting back, do I have any regrets?

I wouldn’t necessarily call them regrets but probably things that I could have done better. Since its in the past, I don’t really fret over it and like to live in the present moment.

“Yesterday is history,
tomorrow is a mystery,
and today is a gift...
that's why they call it present”

― Master Oogway

Enter fullscreen mode Exit fullscreen mode
  • Well one thing I feel I could have changed was try and work in smaller company and teams sooner. I felt those experiences really transformed me and helped me grow as an engineer. You don’t need to be in a big company/enterprise to learn.
  • Also, I feel I should have started writing blogs, listening to podcasts, reading books and trying to attend conferences or give talks sooner. I know it can be intimidating to speak in front of others. I am an introvert at heart but I was able to work around it by taking a chance upon myself

Apart from this, not much!

All the events in my life and career have led me to this point, the people I met with, who taught me valuable lessons, the companies I worked with, which gave me a chance to experience different challenges and much more. I would not change much.

It was a good journey 😄

Post credits 🥳

Btw, A big part of my success was really due to my friends and family , specially shout out to my wife who stood by me through all of these ups and downs and was a constant pillar of support. You rock J!

Also, I believe I got to work under really good leads and managers, They were calm and level headed and gave me lots of good guidance and I honestly learned a lot by just seeing them and talking with them in action on a day by day basis

So whats next for me?

Flash forward and we are here. It’s 2021, I just completed 10 years as a Software tester and automation engineer and am currently a Quality Engineering Manager

So, Where do I see myself going next from here?

TBH, I am not the type of person who would plan out next 5 years of my life. I usually do planning for no more than the couple of quarters and have a general sense of the sort of things I want to try out.

Even though its been 10 years for me, I still feel like a beginner with a hunger to learn and understand how different technologies work and what patterns could scale products as well as teams, so I guess i’m just going to keep moving on in this journey.

Who knows, I might even start a YouTube channel, podcast or write a book 🤷. The future is bright with lots of possibilities, I know for sure that its going to be exciting one.

The end … Or a new beginning 😉!

PS: If you read this whole post, Firstly Thank you for your patience! I hope you could gather something useful out of it. Where are you in your journey? Do share your experience in the comments as well.

As always, Do share this with your friends or colleagues and if you have thoughts or feedback, I’d be more than happy to chat over at twitter or comments. Until next time. Happy Testing and coding.

Top comments (1)

Collapse
 
timyao profile image
Tim Yao

Nice one @automationhacks !

As you mentioned switched from record and playback to script, I have a question hope you can give me some idea. Now, there is a trend for Codeless test automation, how do you think about it? Are people switching from script to record and playback again? What is the difference? Self healing & AI? Thanks.