About
Computer Engineering student specializing in AI & ML with honours in Data Science. Builds end-to-end machine learning systems spanning computer vision, NLP, explainable AI, and cloud deployment, and is preparing a research paper from an AI project.
Experience
-
Web Development InternSkillNova (UptoSkills)Apr 2026 – Jul 2026
Built a backend authentication system (JWT, OTP, 2FA, auth middleware) and refactored React screens to use live API data instead of hardcoded values. Worked in a team-based workflow on production features, following code structure and review practices.
Education
-
2027
B.TechComputer Engineering (Specialization in AI & ML, Honours in Data Science) · 2023 – 2027
Skills
Tools / apps / platforms
Projects
-
Community Digital Library PlatformAzure, GCP, SQL
Deployed a cloud-based library with authentication and a cloud-hosted SQL database on Microsoft Azure and Google Cloud Platform; compared deployment and scalability across both.
-
Farm Produce Traceability PortalSpring Boot, Jenkins, Docker, Puppet
Automated build, test and release with a Jenkins pipeline; 4 Selenium tests act as a quality gate that blocks deployment on failure. Packaged the app as versioned Docker images on Docker Hub and deployed to a clean node with an idempotent Puppet manifest, with health checks and one-command rollback.
-
Multilingual Mental Health Support ChatbotPython, NLP, SQL
Implemented NLP techniques to process multi-language user input and added database logging of conversations; assisted in testing and iterative refinement to improve response accuracy.
-
Insurance Fraud Detection - Text Analysis PipelinePython, DistilBERT, PyTorch
Built a DistilBERT-based NLP pipeline that extracts damage type, severity and location from unstructured claim text and outputs 768-dimensional embeddings. Trained and evaluated the model on GPU, reaching about 99% F1 score.
-
YogaPathshala+ - Real-Time AI Yoga Coaching SystemPython, MediaPipe, OpenCV, Flask
Built a real-time computer vision pipeline using MediaPipe BlazePose and OpenCV; engineered joint-angle features and trained a stacked ensemble (Extra Trees, LightGBM, SVM) to classify pose quality and give live feedback. Applied TreeSHAP for explainability and temporal consistency filtering to stabilize noisy frame-level predictions; integrated the Groq API (LLaMA 3.3) for AI-generated coaching. Served the system through a Flask backend with SQLite session storage; currently preparing a research paper for submission.
🏆 Achievements & awards
-
3rd Prize · 2025
CESA Cognition Clash
-
Participant · 2025
Aavishkar