At L&T Technology Services, I architected a LangGraph-based Agentic AI platform for 5G RedCap UE testing. It plans tests, synthesizes scripts, and validates them, reducing manual test scripting effort by 80%.
I also engineered a RAG pipeline using 5G specifications and historical logs, and integrated Gemma 3 and Azure OpenAI to generate executable PyTest-compatible scripts. I built a Streamlit human-in-the-loop interface for running and verifying complex test workflows.
On Honeywell MAXPRO, I designed Python and VBScript automation frameworks and Jenkins CI/CD pipelines, helping accelerate release cycles and reduce execution time across test cases. For a separate image-captioning project, I developed a VGG16 and LSTM approach that achieved a 0.6038 BLEU-1 score, with findings published in IEEE Xplore.
