.png){kind=logo}
Harsh Tripathi
Mid-level Machine Learning Engineer specializing in LLMs, agentic AI, and risk/fraud modeling
Connect with Harsh
Harsh already has a relationship with Reval, so a warm intro from us gets a much better response than cold outreach.
Recommended
Already have an account?
About
Built and productionized an agentic LLM workflow during a summer internship to transform unstructured clinical reports into analytics-ready structured data, using a LangChain multi-agent design plus an LLM-as-a-judge layer to control quality in a regulated setting. Also has experience orchestrating ML pipelines at Piramal Capital using AWS Step Functions/EventBridge/CloudWatch, with strong emphasis on observability, evaluation rigor, and measurable impact (80–90% reduction in manual data entry).
Experience
Education
Key Strengths
- Built and deployed production agentic AI system converting unstructured medical reports into structured data
- Designed multi-agent LLM architecture (parser/validator/monitor) that generalized across 100+ report formats
- Implemented LLM-as-a-judge quality layer to reduce hallucinations in a regulated healthcare context
- Productionized ML/LLM services with FastAPI + Docker and CI/CD for reproducibility
- Improved inference performance via batching, async endpoints, and token-efficient prompting while keeping latency predictable
- Reduced manual clinical data entry effort by ~80–90%
- Strong ML workflow orchestration on AWS (Step Functions/EventBridge/CloudWatch) with auditable, recoverable pipelines
- Reliability-focused evaluation approach: specs/metrics, unit + scenario tests, offline labeled eval sets, observability, and A/B rollouts
- Effective collaboration with clinicians to translate compliance/workflow needs into schemas, validation rules, and usable prototypes/UI
Discover more candidates like Harsh
Search across thousands of pre-screened, high-quality, high-intent candidates on Reval.
Search Talent