Tech Lead / Architect (hands-on)
Positions : 1
Experience : 8–12 years
Reports to : Founder / Head of Product
Location : Chennai with travel to the pilot hospital
Type : Full-time
About the company and the product
India Health Link is building a new multi-tenant Hospital Management Information System (HMIS) — one platform serving many hospitals — with Artificial Intelligence (AI) features delivered as optional plug-ins on top of a core that works fully without them. It is a greenfield build, and the first pilot hospital goes live in month 6.
Our stack: TypeScript, NestJS (Node.js backend framework), PostgreSQL 16, React, React Native, Amazon Web Services (AWS) — Elastic Container Service (ECS) Fargate, Relational Database Service (RDS), DynamoDB, Simple Storage Service (S3), OpenSearch, Redis, EventBridge — Terraform, GitHub Actions, Playwright, Python for the AI services, and Amazon Bedrock (Mumbai region) for large language models (LLMs).
Why this role exists
This is the person who decides how the platform is put together and then writes the hardest parts of it themselves. Multi-tenancy, the billing engine and the offline synchronization layer are the three places where a wrong early decision costs six months later. This role owns those decisions.
What you will build
- The multi-tenant foundation: tenant isolation, identity and access management (IAM), the role and permission model, and the audit trail.
- The five hubs the whole product hangs off: Master Patient Index (MPI), Encounter / Electronic Medical Record (EMR), Unified Orders, event-driven Charge Capture, and the append-only Stock Ledger.
- The offline synchronization design for tablets used by nurses and at counters, including conflict resolution.
- The plug-in protocol that lets AI services subscribe to events and write suggestions back, without ever writing directly into clinical or financial tables.
Responsibilities
- Own the architecture end to end, and record decisions as Architecture Decision Records the team can still read a year later.
- Write code every week — at minimum the billing engine, the synchronization layer and the tenancy core.
- Set and hold the review bar: schema design, migrations, application programming interface (API) contracts, error handling and tests.
- Break the roadmap into work the pods can run in parallel without blocking each other.
- Mentor two mid-level engineers actively, and make sure no module has a single point of failure after month 3 through pairing rotation.
- Work with the Business Analyst to turn hospital workflows into module specifications.
- Be accountable for the platform passing a third-party Vulnerability Assessment and Penetration Test (VAPT) in month 5.
Requirements
- Shipped at least one multi-tenant Software-as-a-Service (SaaS) product end to end and operated it in production — not only designed one.
- Deep PostgreSQL: indexing, query plans, transactions and isolation levels, partitioning, and migrations against live data with no downtime.
- Strong TypeScript and Node.js. NestJS specifically is a plus, not a requirement.
- Event-driven design in practice: the outbox pattern, idempotency, at-least-once delivery, replay.
- A managed-services mindset — you reach for RDS and Fargate before you reach for a Kubernetes cluster you have to run yourself.
- Has led a team of six or more engineers while still writing code.
Good to have
- Healthcare systems experience: Health Level Seven (HL7), Fast Healthcare Interoperability Resources (FHIR), Digital Imaging and Communications in Medicine (DICOM), ABDM.
- Offline-first or synchronisation-heavy products.
- Ledger, payments or accounting domains.
Your first 90 days
- Month 1 — a walking skeleton in staging: log in, register a patient, book an appointment, bill a consultation. A thin slice through every layer, with continuous integration and deployment (CI/CD) and environments live.
- Month 2 — orders hub, event outbox and audit spine in place; the outpatient department (OPD) flow demoable.
- Month 3 — the demo build (OPD, laboratory, pharmacy and the Clinical Scribe AI plug-in) that the sales conversation runs on.
How we work
Two-week sprints. Trunk-based development with short-lived branches. Code review on every pull request. Database changes only through migrations. Automated tests are mandatory around money, stock and clinical-safety paths
