Murugappan M

SWE 2

Leading design of the fax-to-entry platform that automates pharmacy admin work at MedMe Health (YC W21).

• Led development of an LLM-based extractor that turns unstructured patient questionnaire answers into structured medication data. Working through its long tail of edge cases lifted fax-to-entry accuracy from ~60-65% to 95%+.
• Drove HIPAA compliance: access-logged S3 buckets, PHI/PII scrubbing from logs, and server-side encryption of data at rest.
• Made every service debuggable from one Grafana view, with vendor-agnostic OpenTelemetry traces and metrics across the platform.

SDE 2April 2025 – December 2025

Owned core platform architecture for HyperStart, the company's contract lifecycle management (CLM) product.

• Cut infrastructure spend to a 10% MRR-to-server-cost ratio by profiling usage and reallocating resources over 3 months.
• Built the event-driven job system on AWS SQS and EventBridge that runs all of the platform's async and scheduled work.
• Led the architecture for CRM integrations (Salesforce, HubSpot) that prefill deal data to speed up the client deal-closure pipeline.
• Mentored junior engineers and introduced a standard code-review process to raise code quality.

SDE 1July 2023 – March 2025

Founding engineer on HyperStart CLM, owning features from design through to customer outcome as the product scaled to $300k ARR.

• Architected an LLM-based pipeline that extracts metadata from signed contracts, one of the product's main selling points.
• Cut contract-listing latency to under 5 seconds across 15,000+ records by restructuring responses and tuning queries.
• Led VAPT and static code analysis for SOC 2 compliance, hardening API Gateways and Auto Scaling Groups.
• Built an end-to-end testing pipeline in GitLab CI with Playwright and Docker that gates deployments.
• Benchmarked an open-source PDF converter, then designed and deployed it as a production microservice.
• Built a config-driven UI for stamp-paper procurement, so adding a new Article code is a single JSON change.

SDE InternAugust 2022 – June 2023

Built core ingestion and access-control foundations for the CLM platform.

• Built a Role-Based Access Control (RBAC) system, with database schemas and APIs that manage resource access through user groups.
• Built Google Drive and OneDrive integrations that ingest PDFs for the AI extraction workflows.

R&D Intern (Part-time)

Applied machine learning to anomaly detection for security use cases.

• Built an unsupervised Isolation Forest model that detects anomalous user activity from IP, API URL, and MAC-address signals, an approach that also applies to fraud detection.
• Generated synthetic training datasets and deployed the inference endpoint with Python/Django on Heroku.