Jiya Sharma — Data Scientist

I didn't plan to end up in AI. I graduated with a B.Tech in Mechanical Engineering. But somewhere between thermodynamics and turbines, I kept thinking — what if machines could think? So I made a decision that felt scary and right at the same time: I pivoted completely. Now I'm finishing my M.Tech in AI/ML with a 9.43 CGPA — and I didn't just study AI, I built it from the ground up. I built a GPT transformer from scratch in PyTorch — not fine-tuned, not prompted, actually architected. Quantized it to INT8. Streamed outputs live. Because understanding something deeply enough to rebuild it is the only way I know how to learn. I built multi-agent LangGraph systems that route, plan, write, and judge their own outputs. I built a real-time concentration tracker at 30 FPS with 468 facial landmarks. I built medical imaging AI that doesn't just predict — it explains itself with GradCAM heatmaps and auto-generates clinical PDF reports. My research pushed severe-case disaster response recall from 52% to 86%. That number matters to me. Because AI that fails on the hardest cases isn't good enough. On the data side — I've built ETL pipelines on 180K+ records, Power BI dashboards with live API integration, and ML models hitting 90% AUC in supply chain delay prediction. I write about AI on Medium and LinkedIn because I believe the best engineers can explain what they build to anyone. If you're building something ambitious in AI, GenAI, or data — let's talk. Portfolio: jiyaa.me GitHub: github.com/JIYA1220 Open to: full-time roles, internships, and referrals

Stackforce AI infers this person is a Data Scientist specializing in AI/ML and Computer Vision.

Location: Greater Delhi, India

Experience: 0 mo