Product Engineer • Open Source Builder

Sai Medapati|

Ibuildreliable,observablesystemsatscaleenterpriseplatformsandproduction-gradeopensourcetoolsalike.

CreatorofDayTrail(60+stars,300+downloads,Top3inAIDaily).

Philosophy

Depth over breadth.
Context first.

The best engineering comes from going deep on real problems with real stakes. Years shipping production systems under pressure showed me that elegant code is worthless if it can’t survive traffic, stakeholders, and 3 a.m. incidents.

Every system I build — whether an enterprise integration layer or an open-source AI tool — starts with the actual person who depends on it. Context first. Then technology.

0+★
DayTrail stars (Top 3 Silver)
0–9
engineers led / mentored
0
npm packages published
Past Work

Production systems at scale

Six years owning mission-critical workflows inside Medics HIMS — a hospital information management platform running in real operational environments. Three areas where the engineering decisions actually mattered.

Hospital operations floor
Hospital Finance

Money moving through a hospital, correctly

The problem

Billing, collections, receipts, claims, settlements, credit bills, advances, and payer mapping all touch the same transaction — across patients, payers, departments, facilities, and corporate entities. A rounding error or a mis-mapped payer is not a display bug; it is a reconciliation problem someone has to unpick by hand.

The call

Owned these workflows end-to-end rather than by module, so the finance team could trace any figure back to the event that produced it. Optimised for auditability over convenience — more write-once records, fewer in-place edits.

BillingCollectionsClaimsSettlementsPayer mappingAdvances
Network infrastructure rack
Enterprise Integrations

Ten external systems, none of which agree

The problem

ERP, LIS, CRM, analytics, government and partner systems each expect their own payloads, auth, and delivery guarantees. ABHA, SAP, Oracle, NetSuite, Zoho, Tally, LeadSquared, Business Central — every one a different contract, and every one able to be down when yours is not.

The call

REST and webhook flows built with retries, reconciliation paths, operational logging, and failure visibility as first-class features rather than afterthoughts. The premise: partner outages are normal, so the system has to make them visible and recoverable instead of silent.

ABHASAPOracleNetSuiteZohoTallyLeadSquaredBusiness Central
Hospital corridor
Reliability & Team

The part that happens after go-live

The problem

A hospital platform does not get a maintenance window because it is inconvenient. Production debugging, MongoDB and MySQL analysis, integration triage, and incident coordination are the actual job once a release lands.

The call

Led and mentored a 6–9 person cross-functional team through UAT, rollout planning, and post-go-live stabilisation — and wrote the architecture notes, SOPs, and implementation guidance that let other people handle incidents without me.

Incident coordinationProduction debuggingUAT & rolloutMentoringSOPs
Key Technologies
Node.js, TypeScript, React, Angular, MongoDB, MySQL, Docker, Event-driven systems
What I got really good at
Event architectures • Integrations & reliability • Shipping under real constraints
Healthcare SaaSFinancial SystemsEvent-Driven ArchitectureAudit-FirstEnterprise IntegrationsLocal-FirstProduction ReliabilityOpen SourceHealthcare SaaSFinancial SystemsEvent-Driven ArchitectureAudit-FirstEnterprise IntegrationsLocal-FirstProduction ReliabilityOpen Source
Medics Event Gateway

Enterprise systems connected through auditable event flow

Architected a centralized event-driven integration layer for an enterprise healthcare platform, replacing limited Mirth-based workflows with observable, retryable, auditable webhook delivery. It now handles 50+ business events and 20k+ real-time transactions/day across ERP, analytics, government, lab, pharmacy, CRM, and finance systems.

The call that mattered: build one reusable delivery platform rather than absorb every partner’s requirements as repeated one-off changes to core HIMS transaction logic. Event delivery became its own concern — with multi-tenant context, partner-specific routing, and failures that surface instead of vanishing.

50+ business events
20k+ transactions/day
10+ external systems
Mirth replacement
Retry + audit + recovery
Transformation
Before
Limited debugging visibility in Mirth-based flows
After
Custom event platform with delivery tracking, retries, audit logs, and manual recovery
Flow
retry + backoffFailed delivery returning for retryCaptured50+ eventsNormalizedschema mapDelivered10+ systemsAuditedfull trail
Connected Systems
OracleSAPTallyZohoLeadSquaredBusiness CentralABHALISPharmacyCRMAnalytics
Applied AI

AI as controlled, auditable infrastructure

Not chat wrappers or demo magic. I build AI workflows around grounded data, structured tool calls, approval checkpoints, and audit trails — for healthcare and engineering environments where wrong answers have real consequences.

AI Pipeline
01
Data Stream
02
Model Context
03
Tool Calls
04
Human Approval
05
Operational Output
IVF Treatment Intelligence
Predictive support for treatment planning and trigger-dose estimation, designed for doctor-reviewed decision support.
HIMS LLM Gateway
Controlled LLM access layer for hospital product surfaces, with grounded context, tool boundaries, audit logs, and approval checkpoints.
MCP Tooling
Structured channels between assistants and real systems, enabling controlled tool use instead of uncontrolled automation.
AI Engineering Workflows
Codex, Claude, and Copilot workflows for code review, documentation, migration planning, and delivery acceleration.
Clinical Workflow AI
Decision-support prototypes for specialty workflows, built around review-first operational safety.
Human Approval Layer
Approval gates before AI-assisted actions affect real hospital, financial, or operational workflows.
Developer ToolsDayTrail (60+★ Top 3)MCP & Agentic AIRemoteBridgeLocal-First Systemshuly-mcp-sdkPrivacy by Designnode-locksmithDeveloper ToolsDayTrail (60+★ Top 3)MCP & Agentic AIRemoteBridgeLocal-First Systemshuly-mcp-sdkPrivacy by Designnode-locksmith
Open Source & Developer Tools

DayTrail is the flagship. It has earned 60+ stars and 300+ downloads — and just ranked Top 3 Silver in AI Daily on The Capital.

This is the work I’m most proud of. 73 public repositories. Published npm packages. Every tool started as a real friction I felt while shipping production systems or trying to stay in flow as an engineer. The same thinking that goes into reliable enterprise integrations shows up in the open source tools I build.

DayTrail60+★ / 300+ downloads • Top 3 Silver (AI Daily)Live npm package trends belowLongest-running: node-locksmith (since 2023)
GitHub achievements
DayTrail
60+ stars · 5 forks · 300+ downloads
DayTrail — TOP 3 · SILVER
Ranked #2 Silver • AI Daily on The Capital (week of June 8, 2026)

Quietly captures lightweight metadata from every app, browser tab, editor, terminal, and AI tool you use. Reconstructs your entire day with source-backed evidence. Tracks AI tool usage as first-class work. Gives you perfect daily/weekly summaries for standups, invoices, changelogs, and retrospectives. Nudges you back to focus. Recovers context after interruptions. All local, all private — never sends raw activity to the cloud.

What it solves for real users
  • • “Where did my day actually go?” — answered with traceable evidence
  • • Perfect source-backed summaries for standups, client updates, invoices, OSS changelogs
  • • Makes AI tool usage (Claude, Cursor, Copilot, etc.) visible as real work
  • • Gentle focus recovery after meetings, Slack, or context switches
  • • No surveillance. Your data never leaves your machine by default.
Engineering & craft
  • • Tauri + Rust backend — tiny, fast, native. ~9.8 MB Mac app vs typical Electron bloat
  • • Obsessive privacy: screenshots off, clipboard ignored, browser URLs redacted
  • • AI summaries are strictly grounded in captured local facts — no hallucinations about your day
  • • Complete production-quality repo: app + browser extension + VS Code bridge + release engineering + docs + security policy
macOS primary (Windows CI installers ready). Tiny releases: ~9.8 MB macOS DMG, <6 MB Windows. MIT OR Apache-2.0. Full multi-component repo: desktop app + browser extension + VS Code bridge + terminal support + polished release pipeline.
RustTauriReactTypeScriptLocal-FirstPrivacy
The ecosystem
RemoteBridge CLI
5 stars · 2 forks · v0.1.8 (9 releases)

SSH automation, diagnostics, rsync, deployment helpers, and safety gates. Designed as a bridge so AI assistants can safely operate on remote infrastructure.

RustCLIMCPSSHDevEx
Early-stage (first published ~Mar 2026).
huly-mcp-sdk
4 stars · 2 forks · v0.5.6 (12 releases)

Enables AI assistants (Claude Desktop, Cursor, etc.) to perform real CRUD and workflow actions inside a project management platform via JSON-RPC tool calls.

TypeScriptMCPJSON-RPCAgentic AI
12 releases since launch — actively maintained MCP integration.
node-locksmith
3 stars · v1.0.8 (9 releases since Dec 2023)

Lock-file based single-instance execution to prevent duplicate cron jobs, batch processes, and automation scripts from racing in production.

JavaScriptnpmConcurrencyReliability
Longest-running package (~2.5 years of maintenance).
electron-native-speech
8 stars · 1 fork · v0.1.6 (7 releases)

Exposes OS-level SFSpeechRecognizer to desktop JS. Local, low-latency, privacy-first transcription without mandatory cloud roundtrips.

ElectronTypeScriptNativemacOS
macOS-specific (not cross-platform).
HealthOps
1 star · 1 fork · Go / MIT

Health checks, alerting, and operational visibility tooling born directly from production troubleshooting and incident response patterns.

GoObservabilityReliability
Early-stage. Concept and domain connection are the value.
CodePath
2 stars · Rust

Indexing, semantic context, and retrieval concepts to help both humans and AI tools navigate unfamiliar large codebases effectively.

RustAIDeveloper Tools
Get in touch

I build production systems, AI tooling, and open-source products for teams that need reliability.

Open to product engineering leadership, platform infrastructure, applied AI systems, and select collaborations around developer tools, healthcare SaaS, and local-first products.

varaprasadreddy9676@gmail.com
Contact form
Messages are stored privately and emailed fallback remains available.

Anonymous analytics via Microsoft Clarity (recordings, heatmaps) and our event tracker. No PII sold or shared. Public aggregate stats.

Sai MedapatiProduct Engineering Lead · Healthcare SaaS · Open Source

We use Microsoft Clarity for anonymous heatmaps and session recordings, and store aggregate events in Supabase. No personal data is sold or shared. View public stats.

000/ 100
Click or scroll to skip