Aravind Mallemputi

At Mercedes-Benz Research & Development India, I build Generative AI systems that turn complex system and software requirements into executable MATLAB/Simulink models. My AutoModelGen platform uses LLM reasoning, LangGraph multi-agent workflows, and RAG to reduce engineering development lead time by 40%.

I develop Python automation pipelines spanning document ingestion, vector retrieval, agentic reasoning, model generation, validation, and production-ready Simulink exports. I also built an AI-driven ASPICE Compliance Evaluation System that automates evidence discovery, traceability validation, and audit preparation, saving 200+ engineering hours annually.

I work across OpenAI, Gemini, and Claude models, with Dockerized inference services deployed on AWS and integrated into CI/CD quality gates. I focus on reliable enterprise AI through monitoring, evaluation, groundedness checks, and failure-mode analysis.

Beyond enterprise engineering workflows, I built a virtual health assistant with fine-tuned open-source models and a FastAPI backend, reaching 85–90% response relevance accuracy. My M.Tech research also improved hybrid EV energy management by reducing peak battery current stress and extending battery cycle life.