--- title: "Harsh Shelke (@harsh180502) | Hashnode" description: "ML Research Engineer at Gmango AI Health. Read the latest articles by Harsh Shelke on Hashnode." url: "https://hashnode.com/@harsh180502" --- @harsh180502 # Harsh Shelke @harsh180502 · Seattle, USA · Joined November 2026 ML Research Engineer at Gmango AI Health ### [How I Built an LLM-as-a-Judge Eval System for a Medical Voice Agent (First 14 Scores)](https://aiko-realtime-eval.hashnode.dev/llm-judge-eval-system-medical-voice-agent?utm_source=hashnode&utm_medium=feed) How I built an LLM-as-a-judge eval system for Aiko, a medical voice agent: ten cases, 14 graded visits and where the judge went wrong. 18 min read · [aiko-realtime-eval.hashnode.dev](https://aiko-realtime-eval.hashnode.dev/llm-judge-eval-system-medical-voice-agent?utm_source=hashnode&utm_medium=feed) [2](https://hashnode.com/posts/llm-judge-eval-system-medical-voice-agent/6af3ba8a78ba7b79704a7601) [13h ago](https://hashnode.com/posts/llm-judge-eval-system-medical-voice-agent/6af3ba8a78ba7b79704a7601) ### [Aiko Eval System Appendix: The Ten Cases and Every Graded Visit](https://aiko-realtime-eval.hashnode.dev/aiko-eval-system-appendix-ten-cases?utm_source=hashnode&utm_medium=feed) The ten synthetic cold-consult cases, all 14 graded visits, and the checklist and weights behind the Aiko medical voice agent eval system. 20 min read · [aiko-realtime-eval.hashnode.dev](https://aiko-realtime-eval.hashnode.dev/aiko-eval-system-appendix-ten-cases?utm_source=hashnode&utm_medium=feed) [13h ago](https://hashnode.com/posts/aiko-eval-system-appendix-ten-cases/6af3ba73b21c13587ddc1490)