Software Development Engineer · AI/ML Engineer

I build across the full engineering stack — scalable systems, intelligent applications, and production ML integrations. Currently at Juspay Technologies, shipping ML-driven routing and distributed infrastructure across 300M+ daily transactions.

Juspay Technologies — Engineered ML-powered payment routing (multi-armed bandit explore-exploit), Snowflake-style distributed ID generation at 16K+ IDs/ms, ETL pipelines on Redis Streams, and production Haskell backend services. On-call ownership. Real scale.

Barclays — Built DARS, a data lifecycle management platform using dependency-aware topological ordering (Kahn's algorithm) across 50+ SQL Server tables. Cut DB size by 50%, improved query performance by 40%, automated archival workflows end-to-end.

Infosys — Led a 10-member team delivering a GenAI-integrated Flask backend sustaining 500+ req/sec, shipped 1 week ahead of schedule. Improved content discovery accuracy by 35% with generative AI integrations.

Languages   →  Python · JavaScript/TypeScript · Java · C++ · Haskell · C
Frontend    →  React.js · Next.js · Tailwind CSS
Backend     →  Flask · Node.js/Express · SpringBoot
Databases   →  PostgreSQL · MySQL · MongoDB · SQLite · Redis
AI / ML     →  LangChain · FAISS · OpenAI APIs · TensorFlow · PyTorch · HuggingFace
Dev Tools   →  Docker · Git · Grafana · Celery · Jenkins · Jupyter

Lawyer Up — Multi-LLM orchestration layer (Gemini Pro, GPT-4, Claude) with LangChain agent pipelines and dynamic tool routing. 84% domain-specific accuracy across 720+ requests/day, 30% latency reduction.

Ask Your PDF — RAG pipeline with FAISS vector indexing + OpenAI Ada-2 embeddings. 90% retrieval accuracy, 35% faster query turnaround vs. keyword search.

EduMate — Full-stack assessment platform (React + Java/SpringBoot + PostgreSQL) with real-time speech analysis. 85% scoring accuracy, 100+ concurrent assessments. Built at J.P. Morgan Code for Good '24.

Reachify — Influencer-sponsor engagement platform with campaign creation, ad workflows, and price negotiation. Flask + SQLite backend, Bootstrap/Jinja2 frontend.

Federated Learning and its Uses in the Modern Medical Industry
IEEE ICECA 2025 · Co-authored with Dr. Sivashankar G. and Ishita Goel
Privacy-preserving distributed ML in healthcare — data heterogeneity, communication overhead, and security in decentralized model training.

• B.Tech CSE (AI & ML), SRM IST — CGPA 9.8/10
• BS Data Science and Applications, IIT Madras
• LeetCode Knight — 1700+ rating, 700+ problems solved
• Amazon ML Summer School '24 — selected from 80,000+ candidates
• J.P. Morgan Code for Good '24 participant
• Apple & Infosys iOS Developer Program — 1 of ~1,000 selected
• Pre-Placement Offers: Juspay Technologies · Barclays