About Me
Applied ML Engineer focused on practical AI systems, clean data pipelines, and lightweight RAG/LLM integrations. Strong fundamentals — able to implement regression models and simple neural networks from scratch using NumPy. Experienced in turning ideas into working prototypes using Python, FastAPI, Docker, and vector databases.
Built systems like PROD-IQ (structured + vector retrieval reasoning) and TITAN (graph-based impact analysis). I prioritize clarity, reliability, and transparent evaluation instead of exaggerated metrics. My strength is understanding both the math behind models and the engineering needed to ship them.
What I Do
-
Production ML Systems
Building end-to-end ML pipelines from data collection, feature engineering, model training, evaluation, to deployment. LLM fine-tuning, prompt engineering, and reducing hallucination.
-
Compound AI Architectures
Designing systems that combine LLMs with specialized ML models, vector databases, SQL stores, and knowledge graphs. Built Delta State Graph (tracks KNOWN vs. MISSING vs. ASSUMED) and DCBA algorithm (multi-model orchestration).
-
Data Engineering & MLOps
Building robust data pipelines, feature stores, and model monitoring systems. Experience with web scraping, data cleaning, large-scale preprocessing, feature engineering, and preventing data leakage.
Technical Skills
-
Machine Learning & Deep Learning
85%Supervised/unsupervised learning, neural networks.
-
Python & ML Libraries
85%PyTorch (model building, training loops), Scikit-learn (pipelines, preprocessing, model selection)
-
LLM Fine-Tuning & RAG Systems
80%Fine-tuned LLaMA-3.2B with 4-bit quantization (LoRA, QLoRA). Built RAG systems with ChromaDB, vector embeddings, semantic search.
-
Databases & Vector Stores
65%SQL (MySQL, SQLite), Vector Databases (ChromaDB for RAG), Graph Databases (Neo4j).
-
Full-Stack Development
70%Java (Spring Boot, Spring Security, JDBC), Python (FastAPI, Django), React, HTML/CSS/JS, Tailwind CSS.
Key Achievements & Recognition
-
Best Outgoing Student 2025
PPG Institute of Technology
College-wide award for top academic performance (CGPA: 8.23), hackathon wins, and leadership. Selected from entire 2025 graduating batch.
-
TNSCST Scheme - Twice Selected Statewide
Government of Tamil Nadu
Selected twice under Tamil Nadu State Council for Science and Technology scheme. Awarded ₹10,000 government funding for each AI project.
-
NM-AU-TNCPL Hackathon - Top 20 Nationwide
Naan Mudhalvan Initiative
Ranked Top 20 out of 48,700 teams and 478 colleges in statewide hackathon. Awarded a paid internship for innovative AI solution.
-
NGI-TBI Ideathon - 1st Place
Nehru Group of Institutions - Technology Business Incubator
1st place among 100+ teams. Won ₹10,000 cash prize for innovative startup idea. Demonstrated business viability and technical feasibility to panel of judges.
Professional Certifications
-
AWS Educate Introduction to Generative AI - Training Badge
Amazon Web Services (AWS)
View Certificate -
AWS Educate Machine Learning Foundations - Training Badge
Amazon Web Services (AWS)
View Certificate