Final-year B.Tech. student in AI & ML at Pimpri Chinchwad University, Pune — building production-grade AI systems applied to scientific, industrial, medical, and embodied environments.
vedant = {
"role" : "AI/ML Research Engineer",
"university" : "Pimpri Chinchwad University, Pune",
"degree" : "B.Tech. Computer Science (AI & ML) · 2023–27",
"cgpa" : 8.10,
"focus" : [
"Scientific Machine Learning",
"Physical Intelligence & Robotics",
"Agentic AI Systems",
"LLMs & Retrieval-Augmented Generation",
"Medical AI & Clinical Reasoning",
"Production ML Infrastructure",
"Edge AI & Real-Time Inference",
],
"currently" : "Exploring embodied constraints & interaction-driven intelligence",
}
|
|
Ensemble and Hybrid ML Approaches for Renewable Energy Forecasting
AIP Publishing · Scopus Indexed · Best Paper Award — ICCTVB-25, Sanjay Ghodawat University
|
|
|
RecursiveMAS: A Recursive Multi-Agent Latent Coordination Framework for Embodied Task Optimization
Zenodo · 2026
|
|
|
ML Models to Predict Levulinic Acid Production from Sugarcane Bagasse
Zenodo · 2026
|
|
|
Finance-Aware Knowledge Distillation on Consumer Hardware: Achieving Reliable Financial Reasoning with 4GB VRAM
Zenodo · 2026
|
|
|
When Small Samples Mislead: Uncovering Performance Collapse in Financial Language Models Through Progressive Evaluation Scaling
Zenodo · 2026
|
|
|
Machine-Learning-Based Optimization of Decarbonisation Process of Magnesium Extraction from Dolomite
Zenodo · 2026
|
|
|
Small Language Models for Clinical Reasoning and Medical Decision Alignment Models: LLaMA · BioMistral · Med42v2 · Qwen |
|
Oct 2025 — Apr 2026 |
|
| Project | Description | Stack |
|---|---|---|
| 🤖 Developer Code Intelligence Agent | Fully offline AI coding agent (Ollama): autonomous bug detection, patch generation, self-review & test validation · zero API cost | Python Ollama LangChain |
| ⚡ Antigravity Runtime v5.5 | Production AI execution OS: deterministic orchestration, autonomous self-healing, integrity validation & ms-precision audit logging | Python Docker FastAPI |
| 🧠 LLM KV Optimizer | Modular KV Cache Quantization framework — reduces LLM memory footprint by up to 23× on consumer hardware | Python PyTorch CUDA |
| 🦾 NVIDIA PhysicalAI — Robotics | 3-layer LSTM on NVIDIA PhysicalAI dataset predicting robot actions from 15-step observation sequences; end-to-end PyTorch pipeline | PyTorch LSTM NVIDIA |
| 📖 Research Jarvis | Hybrid RAG engine with real-time arXiv monitoring & local doc ingestion; precision-focused answers with exact technical citations | Python FAISS LangChain |
| 🌿 EcoGuard | AI-powered waste management with automated MLOps pipeline for continuous model training and deployment | Python MLOps Docker |
| 📈 Quantum Risk Engine | Regime-aware volatility modeling with Monte Carlo risk simulations; ML-based market state detection | Python FastAPI Streamlit |
| Award | Event | Year |
|---|---|---|
| 🏆 Best Research Paper | ICCTVB-25, Sanjay Ghodawat University (Scopus Indexed) | 2025 |
| 🥇 1st Place | Code4Society Hackathon 2026 | 2026 |
| 🥇 1st Place | CodeApex 24-Hour Hackathon | 2025 |
| 🥉 3rd Place | National DevCraft Hackathon (Team Tech Titans) · Fluxus 2025, IIT Indore | 2025 |
| 🥉 3rd Place | National AI Hackathon · IIT Indore | 2025 |
| 🏓 National Finalist | Table Tennis · IIM Indore | 2025 |
| Certification | Issuer | Year |
|---|---|---|
| Deep Learning Specialization | DeepLearning.AI (Coursera) | 2025 |
| Supervised Machine Learning | Stanford University (Coursera) | 2025 |
| Mathematics for ML & Data Science | DeepLearning.AI (Coursera) | 2025 |
