Ibin Mathew
Open to AI / ML engineering roles

Hi, I'm Ibin — I build agentic AI systems that make it to production.

AI engineer with 3 years of Python software development and 2+ years implementing and industrialising AI solutions — from requirements analysis through to production. Depth in Generative AI (LLMs, RAG, agentic and multi-agent systems), NLP, and MLOps, with hands-on data engineering at a scale of 10 million to 2 billion records.

AI Engineer · Based in Mannheim, Germany

What I do

Focus areas

Where I go deep — from prototype to a system you can measure, monitor and trust.

Generative AI

LLM applications, RAG pipelines, agentic and multi-agent systems (LangGraph, LangChain, MCP), prompt engineering, evaluation and guardrails. Built and industrialised a production agent system end to end.

LangGraphLangChainRAGMCPMulti-agent

Natural Language Processing

Document extraction and structuring from unstructured sources, retrieval over large text corpora, hybrid search, and text classification models.

Document AIHybrid searchEmbeddingsTransformers

MLOps

Evaluation and monitoring pipelines (Langfuse), automated validation, benchmarking and regression detection, containerised deployment with Docker and Kubernetes, and CI/CD.

LangfuseDockerKubernetesCI/CDBenchmarking

Currently

What I'm working on now

Agentic AI Engineering Intern

Oct 2025 – Feb 2026

NEC Laboratories Europe · Heidelberg, Germany

Took technical responsibility for a cross-functional generative-AI initiative from requirements analysis through to production, working with domain experts, engineers and management.

See full experience

Selected work

Featured projects

Patent · Filed, pending

Multi-Agent Orchestration with Memory and Validation

Co-author on a filed patent covering the orchestration of multiple AI agents with shared memory and built-in validation — the architecture behind the production agent system I built at NEC.

Multi-agentLangGraphPatent

PubMed RAG Question-Answering System

A retrieval pipeline over a large biomedical document corpus, with a systematic benchmark of embedding models, vector databases and retrieval strategies. Hybrid search produced the largest gain — not a larger model.

RAGHybrid searchBenchmarkingVector DBs
View on GitHub
M.Sc. Thesis · Ongoing

Simulation-Based Data Generation & ML Model Development

M.Sc. thesis: a physics-based simulation that generates training data, then an ML model trained and iteratively optimised on it — covering the full loop from data generation to evaluation.

SimulationMLPyTorch

Beyond work

The rest of me

Work is the core of this site — but not all of it. A few things I'm building out.

Let's build something.

Whether it's an AI/ML role, a collaboration or a question about my work — I'd love to hear from you.