| 5+ featured projects |
4 research tracks |
AthenaeumAI currently building |
CV + APIs current focus |
VIT Bhopal CSE AI/ML |
I build complete AI systems: data pipelines, models, APIs, and interfaces that hold up outside the notebook.
I'm a Computer Science undergraduate specializing in AI & ML at VIT Bhopal. My work sits at the intersection of applied AI, computer vision, backend engineering, and research-oriented development.
I like building systems where an idea moves from a dataset or notebook into a usable product: engineered features, evaluated models, reliable APIs, persistent storage, and interfaces people can actually use.
| Focus | What I am working on |
|---|---|
| AthenaeumAI | AI-powered study platform that turns source material into quizzes and adaptive learning paths. |
| Backend systems | API design, system design, databases, and distributed systems fundamentals. |
| Computer vision | Shelf monitoring, plant disease detection, multimodal interfaces, and real-world robustness. |
| Research | Geospatial ML, neuro-fuzzy ranking systems, and practical evaluation methods. |
Languages
Backend & Web
AI / ML
Data & Infra
 |
AI learning system that turns notes into quizzes and adaptive study sessions. Built around chunk-based document processing, semantic filtering, Groq-backed generation, MongoDB persistence, and a React dashboard.
|
 |
Geospatial ML platform for prioritizing urban civic complaints. Combines domain-knowledge labels, KNN density features, DBSCAN hotspot detection, and XGBoost classification. Validated on real NYC 311 records with 95.7% accuracy.
|
 |
Single-camera retail shelf monitor without custom training. YOLOv8 detection plus velocity-predicted centroid tracking for shelf-zone monitoring, dwell-time estimation, product-removal detection, and anomaly alerts.
|
 |
Face-gated voice assistant that only listens when it should. Uses OpenCV Haar-cascade face gating, Whisper transcription, and a rule-based intent classifier across five categories with graceful empty-input handling.
|
 |
Computer vision pipeline for wheat, tomato, and cotton disease detection. Combines Segment Anything Model for leaf/stem segmentation with YOLOv11 disease detection trained on Plant Village data.
|
| Work | Signal |
|---|---|
| GA-ET-IVCFS-ANFIS | Neuro-fuzzy product-ranking framework under uncertainty; best model reached R2 = 0.98. |
| Civic complaint prioritization | Hybrid synthetic + NYC 311 evaluation for geospatial civic issue triage. |
| Multimodal AI Assistant | Solo-authored IEEE-style work on real-time interaction through vision and speech. |
| Parkinson's biomarker mining | Voice-feature mining pipeline; best SVM model reached F1 0.935 and ROC-AUC 0.945. |
 |
Currently reading Designing Data-Intensive Applications |
 |
Current rabbit hole Model Context Protocol and production AI tooling |
 |
Engineering bias Complete AI systems over isolated model demos |
Building Software
That Matters