AI Engineer — Intern
Bangalore, India (On-site) | | Immediate Joining
About Eka.care
Eka.care is India's fastest-growing health-tech platform, building the AI stack that powers modern healthcare. We operate at the intersection of clinical workflows, large language models, and India's national digital health mission (ABDM).
Our products — Eka EMR, Eka Scribe, Eka PHR, and the Eka Developer Portal — are used by 55,000+ clinics, 15M+ patients, and thousands of doctors across India every day. At the core is Parrotlet-LLM, our in-house medical AI model that outperforms GPT-4o and Claude on structured clinical note generation.
About the Role
We're looking for an AI Engineer Intern who thrives at the intersection of language model research and production engineering. You'll work on the full lifecycle of AI-powered workflows — from crafting and evaluating prompts to building scalable infrastructure that connects models to the real world.
This is a hands-on role with high visibility. The systems you build will directly shape product quality and user experience for millions of patients and doctors across India.
What You'll Do
Prompt Engineering & LLM Evaluation
- Design, iterate, and systematically evaluate prompts for LLM use cases — reasoning, classification, summarisation, and tool calling
- Build structured evaluation frameworks (evals) to measure prompt quality, reliability, and regression over model updates
- Apply advanced prompting techniques — chain-of-thought, few-shot, structured output, tool use — and document best practices for the team
- Collaborate with product teams to translate requirements into robust, maintainable prompt systems
Workflow & Agentic System Development
- Architect and implement multi-step agentic workflows that coordinate LLM calls, tool use, memory, and external data retrieval
- Build reliable pipelines with appropriate error handling, fallbacks, and observability instrumentation
- Design context management strategies to handle long-horizon tasks within model token constraints
- Integrate Retrieval-Augmented Generation (RAG) and other grounding techniques where appropriate
Requirements
Technical Skills
- Strong Python programming fundamentals
- Familiarity with LLM APIs (OpenAI, Anthropic, Gemini) and prompt engineering concepts
- Exposure to vector databases, embedding models, or RAG pipelines (even academic projects count)
- Basic understanding of REST APIs and software engineering best practices
- Experience with evaluation frameworks or testing methodologies is a plus
Communication Skills (Essential)
- Written Communication — ability to write crisp technical documents, prompt specs, eval reports, and design notes that non-technical stakeholders can understand
- Verbal Communication — ability to articulate model behaviour, trade-offs, and failure modes clearly in team discussions and design reviews
- Structured Thinking — ability to break ambiguous product requirements into well-defined prompt engineering problems and communicate the approach proactively