Pratik Ahir

Designed and built microservice architecture from a monolithic base using Python, FastAPI (Async), PostgreSQL, and MongoDB; boosted deployment agility by 40%. Implemented real-time streaming chunked responses with FastAPI and Redis-backed state updates, improving user experience by 25%. Built intelligent document APIs with OCR and LLM, automating extraction and increasing processing speed 3x. Implemented ETL pipelines with PySpark and Airflow to prepare document and activity data for analytics and model training. Created end-to-end CI/CD pipelines with Docker, Alembic, and enforced coding standards with pre-commit hooks.

Led backend development for BASF Insight360 – a German-based multimodal RAG-based document analytics and chatbot solution. Designed microservices architecture using Python, FastAPI, PostgreSQL, MongoDB, and Redis, boosting deployment agility by ~40%. Implemented real-time streaming responses with FastAPI and Redis-backed state updates, enhancing user experience by ~25%. Built end-to-end ETL pipelines with PySpark and Airflow to prepare document and activity data for analytics and model training. Migrated transcript data to Elasticsearch for faster search and deployed Dockerized services on AWS with Kubernetes. Enforced coding standards with pre-commit hooks.

Designed and built microservice architecture from a monolith using Python, FastAPI (async), PostgreSQL, MongoDB; boosted deployment agility by 40%. Implemented real-time streamed chunked responses with FastAPI and Redis-backed state updates, improving user experience by 25%. Built intelligent document APIs with OCR and LLM to automate extraction and increase processing speed 3x. Implemented ETL pipelines with PySpark and Airflow to prepare document and activity data for analytics and model training. Created end-to-end CI/CD pipeline with Docker, Alembic, and enforced coding standards with pre-commit hooks.