AI/ML ENGINEER — MACHINE LEARNING
Quiet builder. I teach machines to read, reason and reply — with care for the details.
I build AI systems that actually ship — RAG pipelines in production, transformers fine-tuned past 92% accuracy, response times cut by 62%.
Currently pursuing an MS in Data Science at Vellore Institute of Technology, and spending most of it outside the notebook: containerizing models, wiring CI/CD, and putting language models in front of real users.
Production-grade RAG system for legal Q&A. Two-stage pipeline (offline indexing + online inference) with Gemini 2.5 Flash sub-query decomposition, cutting response time by 62% (85s → 32s). Containerized with Docker and deployed on Azure Container Apps via GitHub Actions CI/CD.
End-to-end NLP pipeline for Google Play Store reviews — sentiment detection, multi-label topic classification and AI-driven insights. Fine-tuned RoBERTa to 92% accuracy (0.91 F1); integrated the Gemini API for category-specific recommendations, served via Gradio.
Deep CNN binary classifier trained on ~14K dermoscopic images for melanoma detection. Optimized TensorFlow input pipeline with prefetching and custom batching for stable convergence. Deployed as a real-time Streamlit app for instant medical image inference.
Basically, I make models ship.
I work across AI, ML and NLP — and I care about the part most projects skip: getting models out of notebooks and into the hands of users.
From fine-tuning transformer models to wiring up Docker, CI/CD and cloud deployments, I like owning the whole journey from idea to a URL someone can actually visit. Always learning what the field invents next.
Qualified the national-level postgraduate entrance exam in Data Science & AI with a highly competitive score.
Qualified the highly rigorous national-level joint admission test for M.Sc. programs across top-tier Indian Institutes of Technology (IITs).
Certification course on Large Language Models, covering architecture, training APIs and deployment.
Certification focused on the fundamentals and applications of Generative AI technologies.
Applied machine learning to real-world challenges at the National Datathon held at SRM University.