Principal AI Engineer - Scalable GenAI Systems- Chennai - 0 to 1 building

Hiring for: An AI-driven US healthcare technology company, recognised among the USA's Top 25 fastest-growing healthcare companies.

Role: Principal AI Engineer - Scalable GenAI Systems- Chennai - 0 to 1 building

Positions: 1

Experience: 5 to 10 years

Work Shift: UK Shift 3.30 PM - 12.30 AM (In office, Chennai)

Location(s): Chennai

Type: On-site / Permanent

Salary: Best in Market

We are looking for a hands-on AI Architect who can own the AI thought vision for the Client and personally drive AI-powered features from idea to production.

This is a builder’s role, not an advisory one — you will architect the AI infrastructure, write production code, integrate AI tools across our stack, and design systems that scale with our patient population. You will be the technical owner of every AI capability client ship.

Indicative time split:

• 80% Hands-On Technical Development
• 20% Other (vision, architecture reviews, stakeholder alignment, tooling evaluation).

Key Focus Areas

• AI Thought Vision — set the technical direction for how AI is applied across the Client’s clinical, operational, and patient-facing workflows.
• Build AI Infrastructure — design and implement the foundational platform (model gateways, retrieval pipelines, prompt management, evaluation harness, observability).
• Full Execution / Hands-On — write production code, build prototypes, and ship AI features end-to-end. This is not a hands-off architect role.
• Integrate All AI Tools — bring together LLM APIs, vector stores, agent frameworks, AI voice/transcription, and AI-powered SaaS into a cohesive, well-governed stack.
• Build for Scalability — design systems that scale across thousands of patients and millions of interactions, with cost, latency, accuracy, and safety in mind.

Roles and Responsibilities

• Own the architecture for all AI capabilities — conversational agents, summarization, clinical decision support, workflow automation, voice/transcription, and analytics.
• Take AI ideas from concept to production — prototype, validate, harden, and ship real applications used by patients and clinicians.
• Build the core AI infrastructure: model abstraction layer, prompt versioning, evaluation pipelines, retrieval (RAG) systems, agent orchestration, guardrails, and observability.
• Design and implement RAG pipelines over clinical and operational data with strong relevance, freshness, and privacy controls.
• Build and operate AI agents and tool-use workflows that interact with EHRs, telephony, messaging, and internal systems.
• Define and enforce evaluation methodology — offline eval sets, online metrics, regression detection, and accuracy/safety benchmarks for every AI feature.
• Drive responsible AI practices — PHI handling, prompt-injection defense, hallucination mitigation, audit logging, and HIPAA-aware design.
• Evaluate and integrate the best-of-breed AI tools, models, and platforms (OpenAI, Anthropic, open-source LLMs, vector DBs, agent frameworks, AI voice).
• Optimize for cost and latency — model routing, caching, batching, distillation, and rightsizing.
• Partner with the Technical Manager and engineering team to ship AI features cleanly into the production platform.
• Partner with Product, Clinical, and Operations leaders to identify the highest-value AI opportunities and convert them into shippable applications.
• Mentor engineers on AI engineering practices; raise the team’s overall AI literacy and execution capability.
• Maintain a forward-looking view of the AI landscape and continuously bring proven advances into the product.

Technical Requirements

• Proven hands-on experience taking AI ideas from prototype to real-time, productiongrade applications running at scale.
• Strong programming proficiency in Python and/or JavaScript/TypeScript, with the ability to write and ship production code independently.
• Deep, working knowledge of modern LLM APIs (OpenAI, Anthropic, and open-source models) — including function/tool calling, structured outputs, streaming, and cost/latency optimization.
• Hands-on experience designing and operating RAG systems — chunking, embeddings, vector databases (Pinecone, Weaviate, pgvector, or similar), hybrid retrieval, and reranking.
• Experience building agentic systems with tool use, planning, and multi-step workflows.
• Strong prompt engineering skills, including evaluation, versioning, and continuous improvement.
• Experience with AI evaluation frameworks — golden sets, LLM-as-judge, A/B testing, and regression tracking.
• Solid backend engineering fundamentals — REST APIs, asynchronous processing, queues, observability, and scaling patterns.
• Working knowledge of cloud platforms (AWS, GCP, Azure) and serverless/low-code environments.
• Strong understanding of data privacy, PHI/HIPAA constraints, and responsible AI practices.
• Good to have: experience with AI voice/transcription (Whisper, Deepgram, ElevenLabs, etc.), AI-powered telephony, or conversational AI in regulated domains.
• Good to have: experience fine-tuning, distilling, or self-hosting open-source models.
• Good to have: experience working with healthcare data formats (HL7v2, FHIR, CCDA) and EHR integrations.

Preferred candidate profile

• Demonstrated technical excellence and a strong portfolio of AI applications taken from idea to real-time production — not just prototypes or proofs of concept.
• Years of experience are secondary to proven ability to ship; we are optimizing for builders who can show concrete, scaled outcomes.
• Track record of architecting AI systems that handle real users, real data, and real reliability/cost constraints.
• First-principles thinker who can decompose ambiguous problems, evaluate trade-offs, and pick the simplest workable AI architecture.
• Strong ownership mindset and bias for shipping — comfortable with iteration, measurement, and continuous improvement.
• Excellent written and verbal communication; able to explain AI trade-offs clearly to engineering, clinical, and business stakeholders.
• Good to have: prior experience in healthcare AI, clinical NLP, or AI in regulated domains.
• Good to have: open-source contributions, published work, or talks demonstrating AI engineering depth.
• Willingness to work to align with US business hours.