Software Engineer - Backend & Data Engineering
I'm a backend and data engineer with over 3 years of experience at PixDynamics Pvt. Ltd., where I successfully delivered a 40% throughput increase on a high-volume insurance document automation platform. My work spans building AI-driven orchestration systems and scalable microservices, including projects like Launchbox, a self-hosted PaaS, and Indic-OCR, a multi-language OCR system. I'm focused on deepening my expertise as a backend developer, leveraging my skills in Python and FastAPI to create innovative solutions that drive efficiency and impact.
Open-source platform enabling Git-push deployments with automated Docker containerization, isolated runtime environments, and dynamic reverse-proxy routing via Traefik.
View Source CodeLightweight decorator-based profiler measuring execution time, CPU time, and memory per function; published on PyPI and adopted by multiple open-source contributors.
Dependency-free HTTP/1.0 server in<500 lines: full TCP socket lifecycle, static file serving with MIME detection, SIGINT shutdown β benchmarked faster than Pythonβs http.server.
View Source CodeDeep learning OCR pipeline for Hindi, Malayalam, and Tamil scripts; Streamlit UI with image preprocessing (deskewing, thresholding).
Jul 2025 β Feb 2026
Built a config-driven orchestration engine for VideoKYC onboarding flows, reducing manual configuration effort by βΌ60% across client deployments. Designed real-time form rendering and async validation logic supporting multiple simultaneous onboarding flows. Improved queue throughput and API response time by 35% via optimized task routing and Redis caching. Promoted from Data Engineer after 8 months based on consistent delivery of client-critical systems.
Oct 2023 β Jul 2025
Architected an end-to-end document intelligence pipeline for insurance workflows: a multi-stage retrieval system using BM25 + Sentence Transformers for candidate page retrieval and Cross-Encoder models for reranking, feeding relevant pages into LLMs for structured information extraction. Increased processing throughput by 40% via Celery parallelism, batching, and elimination of blocking I/O. Built fault-tolerant retry logic, dead-letter queues, and webhook callbacks β reduced pipeline failure rate by βΌ25%. Designed scalable multi-tenant architecture enabling new LOB onboarding with zero code changes.
Mar 2023 β Sep 2023
Developed backend APIs and frontend dashboards for Video KYC, ReKYC, and analytics platforms serving two major banks. Optimized OracleDB queries and caching layers, improving ReKYC API response time by 30%. Built real-time WebSocket dashboards using Django Channels for live KYC activity monitoring.
Master of Computer Applications (MCA) β’ 2024 β Present (Expected 2026)
Bachelor of Computer Applications (BCA) β’ 2019 β 2022
PixDynamics
Bluepond AI
PixDynamics
Interested in collaboration or just want to say hello? Feel free to reach out!