I build end-to-end systems that integrate LLMs into real-world workflows. I prefer systems that are observable, predictable, and don't fall apart under load. Currently building at NEWGEN.

Portfolio • LinkedIn • Email

• Languages: TypeScript, Python, SQL
• Frontend: React, Next.js, HTML/CSS
• Backend: Node.js, FastAPI
• Data: PostgreSQL, Pinecone
• AI / LLMs: LangChain, LangGraph, OpenAI, RAG
• Integrations: Twilio, Firebase
• Infrastructure: Turborepo, Docker, Git, Google Cloud

• The Problem: Resolving a delayed shipment involves messy manual coordination—checking an API, making a phone call, and updating Slack.
• The System: A FastAPI backend that handles supply chain exceptions by coordinating actions across external APIs, voice calls, and chat using a state-driven approach.
• The Constraint: Keeping system state consistent during live voice interactions while handling low-latency Twilio streams and unpredictable external API failures.
• The Outcome: Automates routine delay handling across APIs, voice, and chat, reducing manual coordination to only edge-case exceptions.

• The Problem: Extracting line items from varied bank statements and tax returns required heavy manual review because OCR is imperfect.
• The System: A split-pane Next.js UI pairing the original PDF with extracted data forms. A Flask backend calculates a synthetic confidence score for extractions.
• The Constraint: Rendering forms with 1,000+ extracted fields simultaneously crashed the browser. Fixed by implementing UI virtualization to lock the page at 60fps.
• The Outcome: Processing thousands of documents daily, the system entirely skips manual review for extractions scoring above 98%, routing only edge cases to human operators.

Senior Software Engineer — NEWGEN (Oct 2025 – Present)

• Extracted UI pieces into a decoupled React component library to standardize frontend patterns across the platform.
• Rewrote the engineering onboarding documentation and local setup scripts, cutting the time to first commit for new hires from 6 weeks to 3 weeks.

Software Engineer (Jan 2023 – Oct 2025)

• Built a drag-and-drop UI for composing ML pipelines using ReactFlow. Integrated TanStack Query to poll node execution status and stream live logs.
• Reduced latency on heavy API endpoints by 20% by restructuring underlying Postgres queries with more efficient aggregations and adding targeted indexes.

Software Engineering Intern — AISENCE TECHNOLOGIES (Aug 2022 – Dec 2022)

• Built the real-time group chat feature for a commercial real estate platform using WebSockets and Firebase.

National Institute of Technology Manipur
B.Tech in Electronics and Communication Engineering (CGPA: 8.43) • Graduated June 2023