Feb 2023 – Apr 2024 · 1 yr 2 mos · Bengaluru, Karnataka, India · On-site

• Responsible for end to end design, development and technical roadmap for the company.
• Worked on creating a exentisible API schema using EAV (entity attribute value) model which allowed company to seamlessly pivot between different business models without significant change in backend. This was heavily inspired by Reddit's R2 API design -
• https://www.infoq.com/presentations/reddit-architecture-evolution/
• Implemented Semantic Search for friend discovery using vector similarity and NLP. We created embeddings for each profile based on their interests and represented them in a vector space. Using cosine similarity to figure out the nearest neighbours i.e. most relevant people with similar interests. We used gpt-3.5 turbo model to allow users to execute searches in natural language, which would then interface with a langchain based system to interface with the vector DB.
• Created a user engagement system which took inputs from our pixel driven analytics system, to direct user to most relevant content. Personalization was implemented by taking into account signals from user behaviour (likes, dislikes, interests) and what is trending on the platform (based on hashtags, posts), helping in generate relevant push notifications which had a click through rate of >50% among an audience of ~10000 users.
• Implemented with help of rabbitmq, firebase and in-house analytics system on airflow.

May 2019 – Jan 2023 · 3 yrs 8 mos · Bengaluru Area, India

• My team develops customer facing APIs and search/suggest infrastructure for Bloomreach Search & Merchandising.
• Led migration to microservices architecture from monolithic setup, without downtime. Streamlined release processs, innovated new CI/CD pipelines for build and releases
• Was responsible for Python3 migration along with modernising the entire stack in terms of monitoring, logging and security. Used prometheus, grafana, k8s, efk stack.
• Cross cloud migration, worked on setting up services on Google Cloud to open more deals with customers. This allowed Bloomreach search to be available on both AWS & GCP. Worked on internal solrcloud service to be able to manage clusters in a cloud agnostic way, this was unique to e-comm industry.
• Reworked the backend pipelines to move from batch processing to realtime using airflow, this allowed the customer changes to be reflected on API in near real time.
• Worked alongside Product team in conceptualizing and developing content search and content suggest products, which enabled indexing of schema-free text articles. Was responsible for end-to-end delivery of this project
• Worked with team in adopting API fairness and rate limiting using istio and envoy proxy. Along with using GitOps (argocd) to enable continuous delivery of changes.
• Tech: Python, Shell, Java, Scala, Solr, Airflow, Django, K8s, Docker, AWS, GCP, Elasticsearch

May 2018 – Jul 2018 · 2 mos · Bengaluru Area, India

• Worked on BloomReach auto suggest service (gives text suggestions based on key presses). I ported the service to python3 and deployed it as a microservice after refactoring it out from a monolithic codebase.
• Tech- Docker, Python, Django, Bash, AWS

Jan 2017 – Feb 2017 · 1 mo · Hyderabad Area, India

• 1. Created a test suite for the application, improved test coverage by 40%
• 2. Rewrote APIs and worked on code refactoring, made the code more modular
• 3. Integrated services like Office365, Gmail, Nylas
• 4. Worked on dockerizing the existing django application using docker, gunicorn, nginx, bash.
• 5. Also worked on optimizing database(postgresql) queries and analyzing the runtime of multiple database queries.
• Tech Stack- Django, ReactJS, Docker, AWS, Python unittest, testcoverage, Postgresql

Apr 2016 – Aug 2016 · 4 mos · Bhopal Area, India

• Working on a computational genomics web application SegAnnDB. My work is mainly focused on using improving the chromosome viewer and adding new features. SegAnnDB is the cutting edge application for chromosome and copy number analysis, it achieves this by using ML and user defined annotations.
• Technologies Used - Python, Javascript, Bash
• Frameworks - D3.js, jQuery, Pyramid, Berkeley DB
• Project Demo- https://www.youtube.com/watch?v=cXRxkDfHjtA
• Github Repo - https://github.com/abstatic/SegAnnDB
• Bioinformatics Research Paper (Introduction about SegAnnDB) - http://bioinformatics.oxfordjournals.org/content/early/2014/02/03/bioinformatics.btu072.full.pdf

Oct 2015 – Mar 2016 · 5 mos · Bhopal Area, India

• Did setup for Continuous Integration (Using Git, Jenkins, Gradle, Artifactory) for the team. Routine maintenance and backup of our VPS . Setting up team collaboration via git.
• Was also proactively involved in development and was responsible for Backend programming including but not limited to RESTful web services (Jersey) , Caching (EhCache, JCS) and setting up payment services using Stripe.