Open to opportunities

lalitha ravi

@lalitharavi

Message

AI Engineer building applied LLM and MLOps systems for measurable, real-world impact.

Singapore

Message

What I'm looking for

I’m looking for a team where I can build applied AI/ML and LLM systems end-to-end—data to deployment—with strong engineering practices (MLOps, observability, evaluation) and real-world impact.

I’m an AI Engineer focused on applied ML systems—especially LLM-driven pipelines and MLOps-ready infrastructure. As a Research Associate at NUS, I work with industry and government partners to improve data-centre efficiency and sustainable construction using physics-informed RL for cooling control, document intelligence for embodied-carbon workflows, and BMS-driven energy forecasting.

Recently, I automated CFD simulation workflows in OpenFOAM with parametric geometry and mesh prep, and I built an end-to-end environmental document intelligence pipeline that structures heterogeneous inputs into Json/Markdown for BECC-style downstream use and RAG-ready querying. I also ship production-style ML features—like an LLM-powered conversational agent on AWS with citation-backed RAG—while designing for performance, observability, and reproducible evaluation.

Experience

Work history, roles, and key accomplishments

AI / Data Science Intern

Real Estate Analytics

Jan 2024 - Jun 2024 (5 months)

Shipped an LLM-powered conversational agent on AWS (LangChain, GPT-4, Llama 2, LlamaIndex) enabling natural-language querying over property document corpora with citation-backed answers. Built a RAG pipeline using LlamaIndex and PostgreSQL with Flask REST APIs to ground responses in listing context and reduce hallucinations on proprietary data.

LLM LlamaIndex RAG AWS PostgreSQL Flask REST APIs Prompt Engineering Lang Chain

Software Engineering Intern

GEP Solutions

May 2019 - Jul 2019 (2 months)

Built an MVC .NET application automating CI/CD for containers, microservices, and databases using Azure DevOps, Octopus Deploy, and SonarCloud, cutting manual deployment effort by ~70%. Delivered REST API-based automation and maintained active CI/CD tooling used across the organization.