Skip to content
Back to talent search
RAGGoagentic LLM systems (MCP, tool use)PythonKubernetessearch & retrievalML Product buildingLLM post-training (GRPO/RLHF)vLLMevaluationAWS/GCPranking

Description

Nikhil Kasukurthi is an Applied ML engineer with 8 years of experience building ML Products, specializing in search & retrieval and agentic LLM systems in production. He developed agent-native retrieval (MCP tools, hybrid late-interaction reranking) for grounding LLMs in external data and open-sourced the LLM evaluation library KARMA, adopted by the Gates Foundation. He is currently engaged in ML consulting and independent research on agentic search, including training models with GRPO to plan, issue, and refine search-tool queries. He is proficient in Python and Go, holds EU work authorization, and does not require sponsorship. He prefers remote work but is open to hybrid roles within Germany.