Pranav Prajapati

I’m a GenAI and Machine Learning Engineer focused on turning complex, multi-platform data into accurate, actionable software intelligence. I correlate signals across tools like GitHub, Jira, and Jenkins using embeddings and knowledge graphs to support measurable productivity outcomes (e.g., DORA, SPACE).

At Truxt, I build application and chat-interface harnesses to integrate with tools like openclaw and nemoclaw, and I optimize GenAI performance by caching prompts on VertexAI to reduce both cost and latency. I also integrate enterprise-grade MCP servers (mcp-toolbox, context7) to improve context quality and answer accuracy.

Previously, as a Core-Developer for GSoC, I helped shape future road-maps and maintain sktime by scaling backends for Polars support, maintaining forecasting models on Hugging Face, and implementing the MOIRAI foundation model to enable complex transformer-based time-series modeling through a simple interface. I also contribute as an Open Source Developer—extending Keras ecosystems, migrating architectures for Keras3 compatibility, and shipping open solutions through hackathons and volunteer work.