Gem3s Technologies Pvt. Ltd.•1h ago
LinkedIn
Senior Applied AI Engineer
Chennai, Tamil Nadu, India
Full Time
Senior Level
Full Job Description
Key responsibilities:
- Architecture: design a shared AI platform layer for all four products: model access, prompt and version management, retrieval, tool calling, logging, and cost tracking.
- Model selection: choose and benchmark LLM, speech-to-text, and text-to-speech providers on accuracy, latency, cost, and data residency in the UAE and KSA.
- Evaluation: build automated test suites (evals) with golden datasets to measure every prompt or model change before release. Track hallucination, omission, and accuracy rates.
- Safety and compliance: implement guardrails, redaction of patient data, audit trails, and human-in-the-loop review in line with HIPAA, DHA, Saudi PDPL, and SOC 2/ISO 27001 controls.
- Hands-on delivery: write production code in Python and TypeScript for agent logic, real-time voice and speech pipelines, and EMR integration.
- Team leadership: break work into tasks for developers, review AI-related pull requests, and coach the team on building with LLMs and AI coding tools.
- Infrastructure: work with DevOps on deployment, monitoring of AI behavior, rollback, and cost controls on Azure.
- Clinical partnership: collaborate with clinical specialists to translate medical requirements into test cases and rules for agents.
- Customer-facing support: participate in technical conversations with hospital clients and regulators about AI validation and functionality.
Must-have qualifications:
- 6+ years of professional software engineering, including ownership of systems in production.
- 2+ years building LLM-powered features that real users rely on in production.
- Strong Python.
- Working knowledge of TypeScript and Node.js.
- Hands-on experience with agent design: tool and function calling, structured outputs, multi-step workflows, and handling failures/retries.
- Built retrieval-augmented generation (RAG): chunking, embeddings, vector search (pgvector, Qdrant, Azure AI Search, or similar), and retrieval quality checks.
- Designed and run evaluation pipelines (LLM-as-judge, golden sets, regression tests) and demonstrated output quality measurement.
- Experience with at least one major LLM API (Anthropic, OpenAI, Azure OpenAI, or Google) and its tradeoffs on cost, latency, and context.
- Solid grounding in security and privacy for sensitive data: PII/PHI handling, access control, and audit logging.
- Clear written and spoken English; ability to explain AI behavior to hospital IT teams and non-technical executives.
Nice to have:
- Real-time voice agents (LiveKit, Pipecat, Vapi, Retell, Twilio, or SIP) with latency tuning below 1 second.
- Speech recognition and diarization (Whisper, Deepgram, Azure Speech, AssemblyAI) for Arabic or medical vocabulary.
- Healthcare data standards: FHIR, HL7 v2, ICD-10, CPT, and SNOMED CT.
- Shipped an ambient scribe, clinical documentation, or medical coding product.
- Azure (AKS, Azure OpenAI, AI Search) and observability tools for LLMs (Langfuse, LangSmith, Arize, or similar).
- MCP servers or agent frameworks (Claude Agent SDK, LangGraph).
- Experience leading or mentoring a small engineering team.
- Familiarity with GCC healthcare regulations (DHA, DoH Abu Dhabi, Saudi MOH, NPHIES).
Company
Gem3s Technologies Pvt. Ltd.
GEM3S (Global Excellence in Modern Secure Software) is a trusted technology consulting and software engineering company established in 2009. We deliver secure, intelligent, and scalable digital soluti...
Chennai, Tamil Nadu, India
Posted on LinkedIn