Home

SELECTED WORK

Applied AI built around real problems.

A collection of practical work in machine learning, deep learning, NLP, Generative AI, and LLM-powered applications. Python backend engineering, data systems, and deployment practices support the journey from model experimentation to useful AI applications. Detailed write-ups focus on architecture, trade-offs, evaluation, and lessons learned.

02

ENGINEERING PROJECT

Semantic Resume–JD Matching

A semantic matching workflow that represents resume and job-description text with transformer embeddings and compares relevance using cosine similarity.

Sentence Transformersall-mpnet-base-v2EmbeddingsCosine SimilarityPython
03

ENGINEERING PROJECT

LLM-Powered Resume Analysis

LLM-powered workflows for resume summarization, categorization, and structured information extraction using instruction-following language models.

LLMsMistralNuExtract3Prompt EngineeringFastAPI
04

ENGINEERING PROJECT

Asynchronous ML Inference Pipeline

An AI model-serving workflow exploring vLLM, GPU memory constraints, controlled inference concurrency, and reliable LLM execution through supporting backend services.

vLLMNuExtract3LLM InferenceGPU OptimizationFastAPIDocker
05

ENGINEERING PROJECT

Hela — Portfolio RAG Assistant

A grounded portfolio assistant that retrieves from a curated professional knowledge base and streams responses with supporting sources.

RAGEmbeddingsHybrid RetrievalFastAPI

More technical walkthroughs are coming

I am adding detailed architecture notes and evaluation evidence progressively. Public case studies avoid private employer code, candidate data, and confidential implementation details.

Discuss a project