About Me
Backend Engineer & AI Infrastructure Specialist
Hi, I'm Nikhil – a Full-stack and AI engineer with 6+ years of experience building scalable, distributed systems and AI-powered applications across fintech, media, research, and health tech. I started in frontend/mobile development, later architected microservices and event-driven systems, and now specialize in AI/ML integrations, RAG pipelines, and cloud-native infrastructure.
Known for taking full ownership from concept to production, delivering measurable outcomes, and collaborating with cross-functional teams. Strong expertise in Python (FastAPI, Django), Node.js, distributed systems, PostgreSQL, AWS/GCP, WebRTC, and multi-agent AI architectures. Passionate about solving complex, ambiguous technical problems with scalable systems and intelligent automation.
My Professional Journey
AI Solutions Consultant — Backend & AI Infrastructure
Independent
- •Architected a voice-based document intelligence platform using WebRTC, LiveKit, FastAPI and PostgreSQL; implemented a multi-agent RAG pipeline with CrewAI and LlamaIndex that improved search relevance by 40%, supported 1,000+ concurrent sessions with sub-50 ms latency and 95%+ transcription accuracy.
- •Designed multi-agent LLM reasoning pipelines (CrewAI + LlamaIndex) with a three-tier RAG approach (local/document/global embeddings) powering semantic question generation and context-aware inference.
- •Built continuous streaming infrastructure for STT/TTS (Deepgram Nova-3, Cartesia Sonic) and model inference in production with real-time WebRTC voice UI.
- •Led end-to-end solution from prototyping to production on GCP, leveraging Docker, Kubernetes, Qdrant vector database, and introduced comprehensive monitoring (Prometheus/Grafana) and CI/CD pipelines; collaborated with clients to prioritize features and mentored junior engineers in both front- and back-end best practices.
Full-Stack + AI Developer
Northeastern University
- •Designed a FastAPI + PostgreSQL pipeline enabling semantic search across 10K+ biomedical articles; integrated vector embeddings and search scoring to surface relevant literature for biomedical research.
- •Built AWS Batch cloud pipeline to execute large-scale simulations; automated job orchestration and resource provisioning, reducing compute costs by 40%.
- •Created React + D3.js visualization tools and TensorFlow CNN classifiers for glycan research analysis.
- •Developed a tactile graphics generator and screen-reader plugins to improve STEM accessibility, leveraging React and OpenCV.
Software Engineer
Times Internet
- •Scaled backend services (Spring Boot + Redis) to handle 8.4M daily requests and supported 120K+ subscribers, contributing to $150M+ annual revenue for the TOI+ subscription platform.
- •Designed a Kafka-based personalization and recommendations pipeline that delivered personalized news feeds; achieved 9.7% CTR increase for premium users.
- •Led migration of 70+ city sections into React micro-frontends, improving page load times and raising Lighthouse performance scores to 92/100.
- •Containerized services using Docker and deployed to AWS EKS, reducing infrastructure costs by 35% and simplifying deployments.
Founding Software Engineer
Progcap (Fintech)
- •Designed and built Node.js/Express microservices for real-time loan underwriting; reduced latency from 8.7s to 890ms through caching and optimized database queries.
- •Implemented a Kafka event-driven architecture processing 22K transactions per second with exactly-once semantics to ensure financial integrity.
- •Developed credit scoring ML models (XGBoost) that decreased false negatives by 19%, enabling smarter lending decisions.
- •Supported product growth to ₹9,800 Cr+ lending volume, helping secure Series B funding (US$25M).
Software Engineer
Livemedia
- •Built an OCR-based document verification system using Tesseract.js and Python APIs, achieving 92%+ accuracy for identity verification.
- •Developed a React Native offline-first inspection app used for 50K+ monthly site inspections; reduced inspection time from 45 minutes to 22 minutes through local caching and sync logic.
- •Created a React.js + Django claims platform that cut insurance claims processing time by 60%.
Technical Philosophy
Clean Code Advocate
I believe in writing maintainable, well-tested code that clearly communicates intent. My approach focuses on readability, simplicity, and robust architecture to build systems that can evolve with changing requirements.
User-Centered Design
Technology should serve people, not the other way around. I prioritize intuitive interfaces, accessibility, and performance to create applications that feel natural and responsive to users' needs.
AI Augmentation
I see AI not as a replacement for human intelligence, but as a powerful tool to augment our capabilities. My work focuses on creating AI-powered systems that enhance human productivity, creativity, and decision-making.