I build AI and data systems, and the interfaces people use to trust them.

Senior full-stack and Databricks engineer at Scalata AI. Knowledge graphs, RAG, agents, lakehouse pipelines and the React front ends on top.

This is my work as a graph. Drag a node to pull on it, click one to see what it connects to.

Work

  1. An ontology-backed graph of an organisation's documents, systems and people that anyone can question in plain English.

    Knowledge graphs, Ontology management, RAG, LangChain, Python, Databricks, Vector search

  2. Agents that read a new data drop, map it to the target schema and open a reviewable change instead of a broken job.

    Multi-agent systems, LangChain, LangSmith, Databricks, Delta Lake, Unity Catalog, PySpark, Lakehouse

  3. A fine-tuned classifier that sorts incoming contracts, tickets and reports by type and urgency before a person reads them.

    Model training, Python, MLflow, Databricks, LangSmith

  4. A browser editor for classes, properties and relations, with a live graph view and diff before publish.

    React, TypeScript, Next.js, Radix UI and shadcn, Ontology management, Knowledge graphs, Data visualisation

  5. A small Rust HNSW index with Python bindings, used to keep retrieval local during evaluation runs.

    Rust, Vector search, RAG, Python

  6. Describe a form in a sentence and get a working one: Gemini drafts the fields, you edit, share and export responses to Excel.

    Next.js, JavaScript, Tailwind CSS, PostgreSQL, Stripe, Gemini, React

  7. A trip planner that pulls flights, hotels, weather and attractions into one itinerary and takes payment with Stripe.

    FastAPI, React, MongoDB, Python, TypeScript, AWS, Stripe, OAuth

  8. An end-to-end shop with accounts, product images on S3 and Stripe checkout.

    Next.js, React, JavaScript, Node.js, MongoDB, Stripe, AWS

  9. List a home, favourite other listings and manage reservations, with images served from Cloudinary.

    React, Tailwind CSS, TypeScript, MongoDB, Cloudinary

  10. A Spring MVC REST service with full CRUD, containerised, and covered by unit and integration tests.

    Java, Spring Boot, Docker, PostgreSQL, CI/CD

Open source

  • Added scroll-to-top to the device preview so long pages are quicker to move through.
  • Found and fixed several UI bugs in the open-source CRM.
  • Rebuilt a radio group component on top of the current Radix UI primitive.

Experience

  1. Aug 2024 – now

    Senior Full Stack and Databricks Engineer

    • Design and ship multi-agent pipelines that ingest credit data, align schemas and flag anomalies, using RAG over vector stores and LLM orchestration on Databricks.
    • Build PySpark and Delta Lake pipelines that handle schema evolution and feed unified reporting across credit portfolios.
    • Own the API and real-time layer in FastAPI, Flask and WebSocket, with OAuth sign-in and sessions on Azure App Services and MongoDB.
    • Trace and route agent calls with LangChain, LangSmith and MLflow so each model decision can be inspected.
  2. Aug 2021 – Mar 2022

    Software Developer Intern

    • Built and deployed three client web apps on the MERN stack, cutting manual operations work by about 20%.
    • Stored assets on AWS S3 at 99.99% uptime and wrote unit and integration tests to 85% coverage.
    • Raised organic traffic 10% through technical SEO fixes.
  3. May 2020 – Jul 2020

    Software Analyst Intern

    • Gathered requirements for Project ICE, which rolled SAP finance, HR and materials modules out to more than 500 sites.
    • Wrote functional test cases and cleaned and migrated data for the ERP cut-over.

Education

  • 2024Stevens Institute of Technology
    MS, Computer Software Engineering. Merit graduate scholarship
  • 2022Savitribai Phule Pune University
    BE, Computer Science and Engineering

Skills

Every skill here is a node on the map. Pick one to see where I have used it.

Languages
Frontend
Backend and APIs
AI and knowledge
Data and lakehouse
Cloud and delivery
Domain

Building with AI

I work alongside AI coding agents every day, for design, code and review. These are the tokens I have run through them, synced from my own machines by a usage tracker I built.

Claude

23.2B tokens

  • Opus 5.522B95%
  • Fable 5.1629M2.7%
  • Opus 5438M1.9%
  • Sonnet 5.597.8M0.4%
  • Haiku 4.513.2M<0.1%
  • Haiku 5.5237K<0.1%

Codex

369M tokens

  • GPT-5.6 Sol267M72%
  • GPT-6 Luna55.7M15%
  • GPT-6 Astra44.2M12%
  • GPT-5.6 Terra2.12M0.6%

Last synced Nov 20, 2026. Counts cover the history on my devices only.

Certificates

  • 2024
  • 2024
  • 2023
  • 2020
  • 2019

Contact

Hiring for AI, data or full-stack work? Email is the fastest way to reach me.

sagarhedaoo@gmail.com