PARTHIBAN ● MARIMUTHU
ABOUT / AI & ML ENGINEER

Architecting useful intelligence.

AI/ML Engineer with 5 years of experience taking Generative AI and LLM-powered systems from idea to production. I focus on Agentic AI, RAG pipelines, real-time multimodal systems, and the scalable architectures that make them dependable in regulated environments.

BASED IN
Chennai, India
CURRENT ROLE
Senior Software Engineer — AI/ML (R&D)
EDUCATION
M.Tech, AI & ML — BITS Pilani
(01) / TOOLKIT

Skills

Generative AI

OpenAI, Claude, Gemini, RAG, GraphRAG, Agentic AI, multi-agent systems, prompt engineering, LLM evaluation, multimodal AI, tool calling, structured outputs, MCP.

Knowledge & Data

Knowledge graphs, Neo4j, ArangoDB, Qdrant, vector databases, embeddings, vector search, reranking, document chunking, Redis.

Engineering

Python, SQL, FastAPI, Docker, LangChain, LangGraph, document intelligence, CI/CD, production monitoring.

Cloud & Governance

Azure AI, AWS, GCP, model deployment, AI security, data governance, multi-tenant application architecture.

(02) / EXPERIENCE

Career

OCT 2024 — PRESENT

Senior Software Engineer — AI/ML (R&D), ValGenesis

Leading enterprise GenAI architecture for automated compliance documentation and production-ready AI platforms. Designed graph-based knowledge systems, a 6-stage RAG architecture, multi-tenant FastAPI services, and asynchronous real-time generation pipelines. Developed a Smart Anomaly Detector adopted by five life-sciences enterprise clients.

MAY 2021 — OCT 2024

Machine Learning Engineer — SDE II, Seewise.ai

Built and deployed industrial computer-vision solutions, including a 9-camera galvanizing deployment, downtime monitoring, and vision security systems using OpenCV, PyTorch, Docker, and IoT-triggered alerts.

(03) / SELECTED BUILDS

Projects

GENAI PRODUCT

A+ Studio

Product-listing asset generator producing Amazon A+ imagery and marketing copy, with brand-asset matching and LLM transcreation across 11 EU languages.

REAL-TIME AGENTS

InterVU

AI interview platform with Gemini Live, audio-video streaming, WebSockets, VAD, and multimodal evaluation feedback.

STRUCTURED GENERATION

SVG-Mint

Natural-language-to-SVG system with semantic generation and validation for editable, reusable vector code.

VOICE AI

OExpenso

Voice-first expense tracker turning speech into structured financial data with real-time extraction and categorization.

LET’S CONNECT

Let’s build an AI system people can rely on.

PARTHI.MARIMUTHU10@GMAIL.COM →