go back

my professional landscape


MY WORK & PUBLICATIONS / COMPANIES, I HAVE WORKED WITH / COURSEWARES / BLOG


Personal Projects Relevant Links and Descriptions
  • Coming soon

hyperverge

present

Full Stack (Period.)

An awesome set was awaited. I have been given the priviledge to solve problems of one of the most critical thinking problems here. From designing the architecting the entire pipeline, I have been visiting many business partners to understand their business problems and help setup a suitable workflow where our offerings can fit in and suit them. And helping them grow their and ours business. Going along with the agile product thinking and architecting them has been one of the best thing here. I would categorise the work here as High Engineering and Customer Success.

High Engineering
  • Architecting and developing end to end software and infrastructure that can support configurabe database schema.
  • Worked with GPU technologies. Wrote a CUDA based DB simulations from scratch to support extremely low latency and higher throughput with minimal response time.
  • Designing system to support "Double-Digits-Millions" of identity supports.
  • System design and review to comply with GDPR and various other agreements with clients / 3rd party vendors.
  • Metrics visualisations / scalable deployments / alert systems etc.
Customer Success
  • Visit the business heads, understand their requirements, brainstorm, support in designing work flows, architect their needs.
  • Support customers with the integration and business understanding.

snapwiz

nov' 15 - feb' 17


Associate Software Engineer || Software Engineer

I have worked with snapwiz and worked on end to end product development which included android application, image processing and backend APIs development. It has been a great experience for me. I learnt many things right after graduating from college. Soon after joining snapwiz as an initial "Android Focused Developer" (which I learnt as an hobby in my college days), I jumped into many more roles and started taking responsibilities across computer vision (also learnt in college), backend development etc. The research wing has given me a good exposure to modern practices across various domains.

Across multiple domains, below are few of my key contributions to snapwiz:

** AI, Data Science, Crawls:
  • Built a recommendation algorithm based on location filter
    The app, "glider.ai" has a recommendation engine which recommends candidates to jobs and vice versa based on many factors. One of the key factors was location. The design of this recommendation algorithm was to filter intersections based on location and recommend candidates and jobs to each other.

  • Generic crawler to aggregate candidates' public profile information
    While working for the recommendation algorithms, we had to test our engine with many jobs and candidates. Genereting synthetic data was a solution but we wanted to test with the real data. So, we created a generic crawler that can extract jobs and candidates information from multiple platforms.

  • Image Processing using OpenCV
    One of the interesting tasks, we were caught up was about finding open and closed patches from certain types of images. So, working back on C++ after about 4+ years was a fun and refreshing. Apart from this, I also created a native application for android.

** Architect Design:
  • Solution for dynamic UI rendering which is solely controlled by web service call
    While I was working for an android application, there was an idea that the android application home activity will have dynamic UIs which can be changed on the fly as well. So, architecting this design where tthis can be controlled from the server was relatively a small design but I enjoyed it.

  • Solution for Graph Data Structure and Finite Automata based pagination design
    That's a fancy name. I know. But, the usual pagination based on comparing the inputs from the users and navigating was actually very costly. I was able to decreas the conditional load time for upcoming screens by over 41%. And it worked like a charm.
** Docker & Bash:
When we learnt about Docker and it power, we became excited. So, the research wing wanted me to explore that and migrate the existing projects to it. More than that, there were many small modules that were made in the main app, where some sort of virtualisation was needed while assessing the candidates in the glider platform. Few pointers from Docker:
  • Docker image management simplification with proper hierarchy levels
  • Docker migration for multiple project modules along with their performance profiling and performance testing
  • Multiple shell scripting to automate several repetitive tasks in cloud servers and docker containers
** Brain Storming:
  • Authored many brain storming and complex puzzles & projects in Python and Java8
    Not only that, writing unit tests for these puzzles for the first time was fun. This exercise brought out the old memories from my college.
** Applications Development:
  • Holistic android project development
    I worked on several android applications namely: Glider Jobs (Currently De-Published), Glider Assessment Screen and Edulastic SnapScore.

  • We Service APIs for devices
    The interaction of android application and server was done through socket connections using DDP protocol. I created many server side functions for this.
** Video Processing:
  • Very basic video processing to decrease video file size
    Using multiple compression techniques, changing codecs and frame properties, I was able to decrease certain video sizes so that smaller video can be uploaded from the android / iOS devices in slower 2G / 3G connections.
** Open Source Contributions
  • Worked with an IntelliJ plugin which runs android application in a very light environment
    Emulator roughly takes 3-4 minutes to load and run and android application. The plugin could launch the application within 3 seconds. The downside was that it cannot be used for debugging purpose and native C++ codes cannot be run.

  • Contributed to Meteor DDP Library
    This contribution made social accounts login possible in the Meteor DDP android and iOS library.

freelancing

yes, it happened once

and I am open to get hired there for part time

** AI Gaming's core platform was based out of AWS Cloud. They have been using EC2, Lambda (mostly in Python), DynamoDB and S3. At the time when I started working for Paul from AI Gaming, the idea was to make the most of the backend serverless. And so, for each of the possible RestAPIs, an equivalent Lambda function was created. This freelancing period included below mentioned works:
  • Understanding the business of AI Gaming
  • Transforming their business into logical function mappings
  • Creating AWS Lambda functions
  • Support for executing any source code inside the Lamda using standard inputs
  • Using S3 for dumping many end user's work (for ex: their source code)
  • Integrating DynamoDB for NoSQL based data storage
  • Writing scripts for quick deployment by them
  • Helping them work the same with the ECS equivalent
  • Automated testing and proper knowledge transfer to their internal team members
Paul and Stephen from AI Gaming have been a great help in not just the work, but also explaining the core of their business models. We have been in touch since then and have become friends. :)

ps

disclaimers

  • You might be thinking why I have put a double star (**) while grouping different works?
    Well, the answer is, "just for fun".

  • Many more stuffs are being written and will be coming in the upcoming weeks. Once, they are drafted and made production ready ( :-) :-P ), I will keep removing empty bullets and filling them in with the proper words.

  • I have Zero experience in front end development. So, if you have found any UI issue with this website, please contribute to this repository. Thank You.
Guest Lecture Lovely Professional University
Details coming soon.

deeplearning.ai

since april'18

Deep Learning Specialization, through Coursera.
  • Neural Networks and Deep Learning
  • Improving Deep Neural Networks: Hyperparameter tuning, Regularization and Optimization
  • Structuring Machine Learning Projects
  • Convolutional Neural Networks
  • Sequence Models
Specialization Certificates.

lovely professional university

associated since 2011

Bachelor in Technology Hons.,
Computer Science & Engineering

GPA: 8.08 / 10
Year: 2011 / 2015