Final-year B.E. CSE (AI & ML) student at Chandigarh University (2026, 8.21/10 CGPA).
I build end-to-end AI systems — from research benchmarks to deployed products.
Published researcher · 3 production-deployed AI systems · Open to AI Engineer / ML Engineer roles.
| Project | What it does | Stack | Live |
|---|---|---|---|
| GeoLLM | Natural language → satellite analytics (NDVI, LST, LULC, flood mapping) | FastAPI · GEE · FAISS · Redis · Next.js | demo |
| HallucinationHunter | Training-free hallucination detection for multimodal LLMs (+30.6% POPE accuracy, 4× faster inference) | BLIP-2 · GroundingDINO · CLIP · PyTorch | — |
| Swiggy AI Copilot | Conversational food ordering with natural language cart management and meal planning | FastAPI · Groq (Llama) · React · Vite | demo |
| YOLO-ResNet50 | Rotated bounding box detection for aerial imagery — 7.4% mAP gain on DOTA v1.5 | PyTorch · YOLOv5 · Albumentations | — |
Tiwari, Rana, Kulhari, Upadhyay, Sarmah.
Enhancing YOLO with Rotated Bounding Boxes and ResNet50 Backbone for Accurate Orientation-Aware Object Detection.
ICDAM 2025 · Scopus-indexed · SSRN
Languages · Python · C++ · SQL · JavaScript
AI / DL · CNNs · Vision Transformers · Multimodal LLMs · RAG · Object Detection · Hallucination Detection
Frameworks · PyTorch · OpenCV · HuggingFace · FastAPI
LLM & Retrieval · FAISS · SentenceTransformers · GroundingDINO · BLIP-2 · CLIP · Redis
Infrastructure · Docker · Google Earth Engine · Git · Linux
Open to AI Engineer · ML Engineer · LLM Engineer · Computer Vision Engineer · Applied ML Research Intern roles.
Remote-friendly · Open to relocation.