Aakansha J

I’m a Product Manager with a software engineering background, focused on AI-powered automation, fraud detection, and compliance systems. I’ve shipped products that eliminated ~80% of manual pre-claim review, and I build the evaluation and guardrails that make those systems safe to run at scale.

At AthenaHealth, I led end-to-end delivery across 6 cross-functional teams—discovery through spec-writing, LLM evals, and release. I built an agentic compliance-check engine that routes 100% of visit bills through risk-based screening before claim creation, shrinking a 200-auditor manual review queue to edge-case-only flagging.

I also drove GenAI medical-coding automation from 8% to 92% screening coverage and improved urinalysis automation from 0% to 70%, cutting turnaround ~60%. Alongside that, I expanded RPM-code coverage to 84% and modifier automation to 85%, scaling throughput without proportional increases in manual effort.

My delivery approach is spec-driven development (AIDLC): I replace traditional user stories with structured specs that directly feed AI-assisted code generation, reducing ambiguity and accelerating iteration. I architect LLM evaluation and confidence-thresholding frameworks, design audit-trail and accept/override surfaces for 200+ operations staff, and ensure regulatory defensibility across AI-assisted workflows.