ML Engineer, GenAI developer · Chennai, India

Venkatesh P.

GenAI Developer with a solid ML foundation, building production-ready LLM pipelines, RAG systems, and high-performance vector search. Focused on speed, scalability, and real-world impact.

2,979 QPS. 0.31ms. 55× cache speedup. From medical RAG to semantic search over 42K papers — I own the full stack: embeddings, vector indexes, FastAPI backends, and containerised MLOps on GCP. Biotech background meets production-grade ML.

View Projects → ↓ Resume LinkedIn ↗

Open to full-time ML/GenAI roles

🌐 Open to global relocation

Selected Work

Projects that ship.

01 ★ Flagship Live · Deployed with working

PaperLens

Searches 42,549 arXiv papers by meaning, not keywords — and explains why each result matters. FAISS IVFFlat vector index + LRU cache hits 2,979 QPS at 0.31ms p50; Groq LLaMA-3.3-70B generates research summaries in ~1.2s.

0.31msp50 latency

2,979/squeries/sec

55×cache speedup

~1.2sAI summary

sentence-transformersFAISS IVFFlatFastAPIGroq LLaMA-3StreamlitDockerHuggingFace Spaces

↗ GitHub Live Demo ↗

> init sequence... > loading vectors [dim=384] > index: IVFFlat · corpus: 42,549 papers > cache: 55× speedup on repeated queries > LIVE · p50 = 0.31ms

02 Biology based ML · Research

MS Risk Prediction — GWAS

University-funded ML pipeline for Multiple Sclerosis genetic risk prediction — a problem where most tools are statistical, not predictive. SNP-derived polygenic features into Random Forest + ensemble models at 87.6% accuracy; SHAP interpretability dashboard used actively by the research team. Containerised on GCP with DVC versioning.

87.6%Accuracy

GWASSNP Data

SHAPInterpretable

Random ForestXGBoostSHAPPandasStreamlitGCPDVCDocker

↗ GitHub

03 RAG System

Enterprise Ticket Analysis Bot

RAG pipeline over 70MB+ of enterprise ticket logs — surfaces answers keyword search can't find. PDF → chunking → SentenceTransformer embeddings → ChromaDB → Groq Llama-4 Scout delivers contextual Q&A; REST API via FastAPI, ops dashboard via Streamlit.

ChromaDBGroq Llama-4 ScoutSentenceTransformersFastAPIStreamlit

↗ GitHub

04 LLM + RAG

Well.AI — Medical LLM Assistant

Production medical RAG system shipped at NASO Technologies. Dense embeddings + structured retrieval for health domain Q&A; FastAPI microservices on GCP with Supabase session logging for full conversation auditability — every answer traceable to its source chunks.

LangChainRAGFastAPIGCPML ModelsDocker

↗ GitHub

05 Computer Vision

Brain Tumor Classifier

95% accuracy across 4 MRI tumor classes (glioma, meningioma, pituitary, no tumor) — built for portable clinical inference. Transfer learning + augmentation pipeline, Dockerised for single-command deployment in any environment.

TensorFlowCNNTransfer LearningDocker

↗ GitHub

Work History

Where I've shipped.

JAN 2024 — APR 2024 Internship

AI/ML Engineer Intern

NASO Technologies

Shipped Well.AI to production in 4 months — LangChain + HuggingFace RAG pipeline on GCP with Supabase audit logging for every session.
Reduced environment setup from hours to one command by Dockerising all microservices and standardising multi-service deployment configs.
Designed prompt-engineering workflows for medical domain reasoning; measurably improved LLM answer relevance on clinical queries.

APR 2024 — JUL 2024 Research

Machine Learning Researcher

RUSA Project · University of Madras

87.6% accuracy on MS genetic risk prediction — engineered polygenic SNP features from GWAS data; trained Random Forest + ensemble models on high-dimensional genomic input.
Interpretability dashboard with SHAP and polygenic risk scoring adopted by the active research team for ongoing clinical analysis.
Reproducible containerised ML pipeline on GCP with DVC data versioning — experiments tracked and re-runnable by any team member.

Tech Stack

Tools I ship with.

LLM / GenAI

RAG PipelinesFAISSChromaDBLangChainsentence-transformersGroq APIHuggingFacePrompt EngineeringLoRA / QLoRA

ML / Deep Learning

PyTorchTensorFlowscikit-learnXGBoostCNNsTransformersMLflowSHAP

Backend & MLOps

FastAPIDockerDocker ComposeGitHub ActionsGCPAWS EC2 / S3DVCHuggingFace Spaces

Languages & Data

PythonSQLBashPandasNumPyPostgreSQLStreamlit

Education

Master of Computer Applications (AI/ML) — University of Madras, Chennai · 2023–2025
B.Sc. Biotechnology · 2020–2023 — rare interdisciplinary base for genomics ML

Certifications

AWS Data Engineering BootcampUdemy · 2024

Machine Learning A–ZUdemy · 2024

Deep Learning SpecializationCoursera · In Progress

Recognition

Awards & highlights.

🥈

Top 5% — National AI Hackathon

National Finalist out of 500+ participants. Competing at national level as a grad student says more than a certificate.

🔬

RUSA Research Fellow

University-funded research grant — 87.6% accuracy on MS genetic risk prediction. Not a class project. A funded study.

✎️

Technical Writer on Medium

Published: EDA techniques, NLP pipelines, Genomics ML. Writing is how I verify I actually understand something. ↗

🌐

Rotaract VP — 150+ Mentored

Led 5+ AI awareness events. Mentored 150+ students in physics and tech. Communication and leadership, not just code.

🎓

Biotech × ML — Rare Stack

B.Sc. Biotechnology + MCA (AI/ML). Most ML engineers can't read a GWAS study. I can build one and productionise it.

☁️

Deployed on GCP + AWS

AWS Data Engineering Bootcamp + ML A–Z. Production deployments on GCP (internship) and AWS (personal projects).

Get in Touch

Ready to scale
your ML pipelines?
Let's build.

Open to ML Engineer, GenAI developer and Data Scientist roles — Chennai, Hyderabad, remote, or global relocation. Response within 24 hours, guaranteed.

venkateshpvnky9@gmail.com

→

+91 91766 36234

→

linkedin.com/in/venkatesh-ml

→

GitHub

github.com/venkatesh-hyper

→

Medium

medium.com/@venkateshpvnky9

→