Resume | Krishnasai Hemanth Randhi

Profile

Four years taking AI from discovery to production in regulated healthcare. Specs it, builds it, measures it honestly. $500K+ ARR shipped, $5–7M in client savings identified, and one $1M project killed on purpose.

krishnashemanth@gmail.com · linkedin.com/in/krishnashemanth · krishnashemanth.github.io

Experience

Product Owner

Symplr · Enterprise Healthcare GRC SaaS · Bengaluru, IN

Aug 2025 – Present

Shipped an AI quote-extraction tool to production in one quarter as PO and sole builder (Claude on AWS Bedrock, FastAPI, React): quotes per analyst per month up 11.7%, line items per quote up 30% with handling time flat, ~99.9% accuracy validated in UAT.
Won funding over 5 competing initiatives with a WSJF case. Shipped CSV-first to avoid a cross-team write-back dependency, returned "N/A" instead of guessed prices, and kept analyst review on every extraction.
Rebuilt ~60–70% of a 20-year, 366-screen legacy platform as a web app, solo (React/TypeScript, .NET 8). Built a labeled eval harness scoring extraction recall, precision and row-level accuracy.
Built 2 regulated-AI POCs: regulatory-change monitoring (per-source scan time from 2 days to under 1 minute) and clinical-evidence extraction with side-by-side human verification. Stopped the follow-on report-drafting POC for lacking clinical pass criteria.
Identified $5–7M in annual savings for an enterprise hospital client through a clinical AI POC. Held 67 NPS on the spend management product with weekly SLA triage across QA, eng and CS.

Senior Business Analyst · Product

Quantiphi · AI-first digital engineering · Bengaluru, IN

Feb 2022 – Aug 2025

Led discovery for a 0→1 clinical radiology AI platform (FDA SaMD, multi-tenant). 20+ clinician interviews across 15+ hospital systems drove the pivot from speed-first to interpretability-first. $500K+ ARR in 6 months.
Raised NLP accuracy from 68% to 88% by routing inline clinician corrections into prompt tuning within the same sprint (Gemini, RAG, HITL). Mapped FDA SaMD and HIPAA requirements into acceptance criteria.
Designed the structured-data pipeline that eliminated 4,000+ manual hours/year of post-processing across partner sites.
Wrote the kill memo that ended a $1M care-gap AI initiative 2 weeks into discovery: PHI/PII access was blocked by a 6–12 month BAA path, and ROI was negative in every scenario. The team shipped a parallel initiative 3 months later.
Structured iterative A/B tests on 4 checkout friction points for a B2B retail client, with the product team; cart abandonment down 42%.

Capabilities

Product strategy

0→1 discoveryMetric design (North Star, guardrails)WSJF / RICEPRDsA/B testingKill memosGap analysis

AI & regulated SaaS

LLMsRAGEval harnessesHITL designPrompt designStructured outputUnit-cost modelingFDA SaMDHIPAAGDPR

Build

Python (FastAPI)React / TypeScript.NET 8AWS BedrockSQLClaude Code

Tools

PendoJiraAzure DevOpsConfluenceFigma

Education & Certifications

Education. M.Tech, Industrial Biotechnology · NIT Karnataka (NITK Surathkal) · CGPA 8.2
B.Tech, Biotechnology · NIT Andhra Pradesh (NITAP) · CGPA 7.3

Certifications. SAFe 5.0 POPM · Product Management Certificate, Product School · Google Associate Cloud Engineer

Awards. Dream Team, You Make a Difference (Symplr) · Kaizen, SPOT (Quantiphi)

Selected case studies

AI Quote Automation Quotes per analyst up 11.7%/month, ~99.9% accuracy, in production.
Clinical Radiology AI · 0→1 NLP 68 to 88%, $500K+ ARR in 6 months.
The $1M project I killed on purpose Compliance-driven discovery kill memo.
Phoenix: Legacy Modernization Prototype 20-yr, 366-screen legacy rebuilt as an AI-first web app.

All 9 case studies →

Let's build what actually ships.