Projects
Selected systems and applied ML work
A mix of application engineering, model development, data analysis, MLOps, LLMOps, and computer vision projects.
QC-Leap
AI-Powered Manufacturing Quality Control System
GitHubNext.js, TypeScript, Tailwind CSS
- Developed application workflows for automated defect detection using computer vision.
- Focused on integrating ML models into a usable production-style interface.
System Threat Forecaster
System vulnerability classification
GitHubPython, Scikit-learn, Pandas, Matplotlib
- Built a classification model to predict system vulnerabilities.
- Implemented preprocessing, evaluation, and model performance analysis.
Household Services Platform
Role-based services marketplace
GitHubVue.js, JWT, Celery, Redis, Mailhog
- Built a full-stack application with admin, customer, and professional roles.
- Developed APIs, authentication, async task handling, and workflow automation.
Influencer Marketing App
Brand and influencer collaboration backend
GitHubFlask, SQLite, Python
- Designed database schema and REST APIs for campaign and user data handling.
- Built backend flows to connect brands with influencers.
Business Data Analysis
Local dairy shop sales analysis
GitHubPython, Pandas, Excel, Matplotlib, Seaborn, SQLite, Flask
- Analyzed sales and customer data to identify trends and improve inventory decisions.
- Created visualizations to communicate insights clearly.
Multilingual Speech Recognition
Speech-to-text pipeline
GitHubWhisper, Faster-Whisper, Indic tools
- Built multilingual audio transcription workflows with optimized inference.
- Handled environment setup and dependency management for model execution.
Multilingual Sentiment Analysis
LLM-based sentiment classification
GitHubGemma, Prompt Engineering, Python
- Leveraged Gemma for sentiment classification across multilingual datasets.
- Designed prompts and evaluated model outputs for consistency and accuracy.
Denoising and 4x Super-Resolution
Low-light image restoration
GitHubPython, U-Net, PixelShuffle, TTA
- Built an end-to-end restoration pipeline for IIT Madras DLP 2026 NPPE3 on RELLISUR.
- Reduced competition RMSE from 41.8 to 18.7 using SRUNet, multi-task loss, and calibration.
- Used EDA to model non-Gaussian noise and improve validation across lighting conditions.
End-to-End MLOps Pipeline
Cloud-native ML deployment and observability
GitHubDVC, MLflow, FastAPI, Docker, GKE, GitHub Actions, OpenTelemetry
- Orchestrated data versioning, experiment tracking, model registry, and deployment on GCP.
- Automated CI/CD for a Dockerized FastAPI service on GKE with HPA and load testing.
- Integrated MLSecOps, drift monitoring, fairness assessment, SHAP, and LLM guardrails.
RAG Document QnA
Local retrieval-augmented document assistant
GitHubPython, FAISS, Streamlit, Gemini
- Engineered a local RAG system for grounded answers from proprietary documents.
- Optimized CPU-bound semantic search with FAISS flat L2 indexing.
- Built a Streamlit dashboard with PDF ingestion, session state, and constrained prompts.