AI · Data Science · Product Engineering
I’m Gundeep—an AI and data science engineer focused on reliable agents, retrieval systems, model evaluation, and the web products that make them useful.
Based in India · Building globally
Featured case studies
From agentic analytics to in-browser retrieval and model reliability—these projects show how I approach applied AI beyond the demo.
01A human-supervised analyst that turns plain-language questions into validated, read-only SQL or Pandas. Tools are discovered through MCP, while headline figures are independently recomputed before display.
Private client engagement
02Interview preparation grounded in your own documents. Parsing, hybrid search, reranking, and citation checks all run inside the browser.
03An independent evaluation of a CVPR 2025 segmentation model across 499 images, measuring robustness, calibration, speed, and failure modes hidden by averages.
Selected builds
ML-driven text classification workflow for evaluating content authenticity.
35 privacy-minded tools to convert, compress, edit, and merge files in the browser.
Clean notebook-to-PDF conversion with live preview, entirely on-device.
A React and FFmpeg-powered workflow for video, audio, Shorts, and playlists.
Services, pricing, publishing, user accounts, and payments for a digital studio.
A full-stack store with catalogue, cart, wishlist, checkout, accounts, and admin tools.
A privacy-friendly guided reflection on autism traits—designed as a prompt, not a diagnosis.
Move text between devices using a one-time code—no account required.
Healthcare management platform with analytics and secure operational workflows.
Interactive visual explanations for core data structures and algorithms.
WebSocket messaging with presence, responsive conversations, and delivery states.
A focused video-conferencing experience for browser-based collaboration.
A campus community platform connecting students, clubs, and academic resources.
Technical range
I work across the stack so promising models become dependable tools—not isolated notebooks.
Reliable LLM workflows, computer vision evaluation, classical ML, and human-in-the-loop systems.
Grounded answers, measurable retrieval, analytical agents, and defensible data transformations.
Responsive interfaces and production backends that make complex capability feel approachable.
Engineer · Researcher · BuilderAbout me
I’m pursuing a PGP in AI & Data Science at Jio Institute after a B.Tech in Computer Science. My work sits where evaluation, reliable system design, and thoughtful user experience meet.
Through YonoSol, I also design and ship products for real clients. That experience keeps my AI work practical: I care about privacy, failure states, maintainability, and whether the person on the other side can actually use what I build.