Experience
Backend Developer Intern - Engineer's Cradle
Built a data ingestion service in TypeScript (Bun + Hono) integrating 3+ social media REST APIs to aggregate cross-platform analytics into PostgreSQL for dashboard reporting.
Designed normalised relational schemas for OAuth2 token lifecycle, permission scoping, and integration metadata using Kysely, enabling secure multi-account connections across five social platforms.
Implemented OAuth2 authorisation flows with token refresh and permission management across five providers, ensuring secure and reliable third-party API access.
Exposed normalised REST APIs for analytics dashboards and engineered sync logic handling API rate limits, retries, and provider-specific inconsistencies.
Full Stack Developer - NextGen QuickHire
Led backend development of an online assessment platform using Node.js, Express.js, and TypeScript, designing PostgreSQL database schemas and REST API architecture for test creation, submission, evaluation, and results workflows.
Built and tested REST APIs for the complete assessment lifecycle, applying typed validation with Zod and standardised API contracts to reduce integration bugs and ensure consistent workflow processing.
Implemented a payment module for secure transactions and order tracking using Razorpay, enabling real-time payment processing for the e-commerce platform.
Designed modular backend services using PostgreSQL and MongoDB with query optimisation to support extensibility across multiple platform modules.
Projects
Natural language → structured financial workflows via specialized AI agents.
Breakthrough Concept — Google Cloud & Hack2Skill
Multi-agent platform that routes natural-language prompts to specialized agents, Trader, Tax Filing, Debt Optimization, Compliance, through a controlled orchestration layer. Built with a 3-member team; led backend and agent infrastructure.
Key highlights
- •Designed agent routing with intent classification and financial rule validation; permissioned execution pipelines ensure agents only run pre-approved, compliant operations.
- •Built REST APIs for prompt ingestion, agent reasoning, and workflow execution with Firebase Auth, authorization middleware, and per-action audit logging.
- •Deployed FastAPI services and agent microservices independently on Google Cloud Run / Compute Engine via Docker and GitHub Actions CI/CD.
Resume-grounded job content generation via a distributed async AI backend.
Four-service distributed backend: Go API gateway, Python FastAPI, Celery worker, RabbitMQ broker: that orchestrates multi-step AI content generation anchored to the user's resume via a RAG pipeline.
Key highlights
- •AI engine built with LangGraph state graph and Gemini LLM: tool calling, schema-validated outputs, and retry guardrails across multi-step generation flows.
- •Async job lifecycle (accepted → queued → processing → complete/failed) via Celery and RabbitMQ — requests return instantly while agents run in the background.
- •RAG pipeline: PDFs parsed with PyMuPDF, chunks vectorized via FastEmbed, indexed in ChromaDB, semantically retrieved at generation time to ground all outputs in resume content.
Ask questions over any uploaded document using vector search and Gemini.
RAG-based Q&A system that lets users upload documents and query them in natural language, combining ChromaDB semantic retrieval with Gemini reasoning for grounded, multi-turn responses.
Key highlights
- •Ingestion pipeline: upload to S3 via CloudFront → extract text → chunk → embed with BAAI/bge-small-en-v1.5 → index in ChromaDB.
- •Vector similarity search surfaces the most relevant chunks per query, minimizing irrelevant context sent to the LLM.
- •Gemini reasoning layer with multi-turn conversation support and explicit out-of-scope guardrails to prevent hallucination.
Immutable double-entry bookkeeping wallet with idempotent transaction processing.
Backend financial service in Go where every transaction is a permanent debit/credit ledger entry. Balances are derived state, never stored directly, making reconciliation deterministic and history always reconstructible.
Key highlights
- •Double-entry model with SQLC type-safe PostgreSQL queries; balance drift is impossible by design.
- •Idempotent transaction processing via client-supplied request IDs prevents duplicate credits/debits under network retries.
- •JWT-secured REST endpoints with strict input validation and structured error handling; containerized with Docker.
Skills
CorePython, Go, TypeScript, Node.js, FastAPI
FamiliarBun, Hono, Express, Flask
CoreLangChain, LangGraph, RAG Pipelines, Agent Orchestration
FamiliarGoogle ADK, Vector Search, ChromaDB, Embedding Pipelines, MCP
CorePostgreSQL, MongoDB, Redis
FamiliarMySQL, SQLite, Turso, Drizzle, Prisma, SQLC, SQLAlchemy, Kysely
CoreDocker, GCP (Cloud Run, Compute Engine), GitHub Actions
FamiliarAWS (S3, CloudFront), Cloudflare Workers, Kubernetes, Linux, Prometheus, Grafana
CoreSolidity, Foundry, Ethers.js
FamiliarWagmi, Solana Web3.js, SPL-Token
CoreReact, Next.js, TanStack Query, Tailwind CSS
FamiliarReact Native / Expo, Zustand
Achievements
Google Agentic AI Day 2025 - Breakthrough Concept Award
Awarded by Google Cloud and Hack2Skill for presenting an innovative agentic AI solution for Fi Money, demonstrating cutting-edge applications of autonomous AI agents in fintech.
ETHGlobal Delhi 2025
Collaborated on Dagent — a decentralized AI agent marketplace with on-chain staking and rewards — at ETHGlobal Delhi, working alongside leading Web3 developers and industry mentors.
CodeXCaliber – Winner
Won the CodeXCaliber competition at IIIT Nagpur, building an AI-powered productivity solution that leverages machine learning to enhance user workflows.
HackFest 2024 – Finalist
Reached finals with a sustainability-focused solution for real-time monitoring and reduction of industrial water footprint, addressing critical environmental challenges.
Smart India Hackathon – National Winner
Selected as the national winning team for developing a comprehensive digital solution for water footprint measurement, contributing to India's sustainability goals.
Education
Bachelor of Engineering in Computer Engineering
Pursuing B.E. in Computer Engineering. Expected graduation May 2026.