Computer Science and Engineering undergrad at Dwarkadas J. Sanghvi College of Engineering, specializing in Data Science.

I build backend systems that carry real load: market data pipelines feeding live algorithmic strategies, lead intelligence engines that replace manual work outright, and health platforms that turn raw lab reports into something a person can actually read. Most of what I ship is measured — latency, Lighthouse, conversion, coordination hours removed — because unmeasured work is indistinguishable from work that didn't happen.

Currently deep in Java DSA, system design, backend architecture, and generative AI integrations.

Jul 2026 – Present · Remote

Building the broker-facing backend that carries an algorithmic strategy from code to live automated execution, spanning market data ingestion through to real orders placed for real traders.

• Migrated market data pipelines from AWS S3 to EBS-backed storage, cutting read latency on the hot path that live strategies depend on for entry and exit decisions
• Built a strategy SDK so a strategy authored in code deploys straight into automated execution, and debugged the SQS → Lambda → ECS pipeline behind it across queueing, invocation, and container rollout

May 2026 – Jul 2026 · Mumbai, Maharashtra

• Shipped a full-stack lead deduplication and filtering tool in React.js and Node.js/Express, letting business teams strip existing clients out of incoming marketing lead lists before a single sales call was made
• Engineered the backend matching layer and tuned the SQL queries behind it to cross-reference every incoming lead batch against millions of existing customer records in one automated pass
• Integrated third-party data validation APIs to auto-verify customer PAN and Aadhaar details at ingestion, rejecting fake and malformed entries before they could reach the CRM

Jan 2026 – Jun 2026 · Melbourne, Australia · Remote

• Replaced manual patient discovery with a TypeScript lead-discovery agent monitoring public Reddit, Facebook, and Whirlpool posts, polling every 15–20 minutes to stay inside rate limits and harvesting 1,200+ raw interactions, fingerprint-deduplicated into PostgreSQL
• Integrated the Facebook Graph API, Apify, and SERP API to surface occupational-therapy intent across Melbourne and Adelaide, ranked by multi-signal intent scoring and a 4-tier priority classifier
• Delivered a role-based React admin dashboard for lead prioritization, deduplication, filtering, and outreach tracking, cutting therapist coordination work by 50%

Feb 2026 – Apr 2026 · Mumbai, Maharashtra

• Owned full-stack delivery end to end, architecting the REST APIs and restructuring the database query paths behind them so the frontend read what it needed in one round trip instead of several
• Enforced testing and debug pipelines across a hybrid development environment, holding zero-downtime deployments and stable interfaces under load through every release

Sep 2025 – Jan 2026 · Chennai, Tamil Nadu · Remote

• Removed performance bottlenecks across high-traffic Next.js pages using component-level code splitting and lazy hydration, lifting Lighthouse scores by 40%+
• Eliminated Cumulative Layout Shift regressions by 60%+ across every breakpoint, holding the interface stable under load on both desktop and mobile
• Shipped a reusable WCAG 2.1 AA-compliant React design system on a structured weekly release cadence, compressing Figma-to-production turnaround by 30% and lifting session engagement 35%+

Jun 2025 – Sep 2025 · Mumbai, Maharashtra

• Removed the human from lead qualification with a real-time AI voice agent on GPT-4o Realtime and FastAPI WebSockets, automating STT/TTS intent routing, scheduling, and follow-up triggers at sub-300ms latency and cutting manual effort 60%
• Engineered a CRM lead capture pipeline enriching every live inbound call into PostgreSQL with algorithmic intent scores and structured call summaries
• Rebuilt the landing pages with lazy hydration and asset compression, lifting Lighthouse scores by 45% and inbound conversion rates by 30%

React TypeScript FastAPI PostgreSQL ChromaDB Neo4j PaddleOCR OpenRouter ElevenLabs Gemini Three.js Docker

A health companion that reads your medical reports and tells you what actually needs attention.

• Physics Twin scoring engine on PaddleOCR + pdfplumber ingests uploaded medical reports and computes 0–100 Health Indices across 6 organ systems through clinical normalization
• Auto-detects 10 conditions with severity mapping, turning a wall of raw lab values into a ranked, plain-language read
• Streams SVG body-impact overlays onto a Three.js 3D viewer via live SSE telemetry at 2-second intervals, so a flagged marker lights up the organ system it belongs to
• Mem0 + Graphiti persistent knowledge graph on Neo4j enables cross-session contradiction detection across a patient history spanning many reports
• Voice agent combining Web Speech STT, Gemini-powered RAG, and ElevenLabs TTS with emergency detection
• Medicine Finder querying a 600,000-entry Jan Aushadhi drug catalog with live pharmacy lookup

Solidity 0.8.26 Next.js 15 FastAPI Python Lit Protocol v7 IPFS Ponder v7 OpenZeppelin v5

Model ownership on-chain: list a model, control who runs it, earn on every call.

• 6 Solidity 0.8.26 contracts deployed to Ethereum Sepolia
• ERC-721 model NFTs with 10-payee revenue splits, per-call escrow with pull-payment settlement, EWMA staked node selection, ERC-5192 soulbound reputation badges
• Real HuggingFace pipeline execution on staked nodes with Lit Protocol v7 per-token access control and IPFS encrypted weight delivery — model files stay private while payments and ownership stay publicly verifiable
• FastAPI orchestrator with SIWE wallet auth and nonce-serialized settlement signing to keep concurrent payouts from racing
• Ponder v7 GraphQL indexer plus a LangChain wrapper published to PyPI, so agents can call marketplace models as ordinary tools
• Full Next.js 15 App Router frontend across 20+ pages: wallet login, model discovery, upload, purchase, ownership dashboard

Languages

Frameworks

Backend & Tools

Cloud & Databases

AI & ML

_{Also: REST APIs · JWT · AWS Lambda, EC2, S3 · ChromaDB · RAG · LangGraph · Multi-Agent Systems}

Finalist at 8+ national hackathons overall.

Creatives Co-Committee Member · DJS S4DS · Sep 2025 – Apr 2026 UI/UX, posters, and graphics for technical events and society initiatives.

Technical Co-Committee Member · DJSCE Trinity · Aug 2025 – Jan 2026 Ran technical events and contributed to the official fest website; organized hackathons and coding competitions end to end.