Artificial Intelligence Engineer

AI Engineer

Ekko | Full-time | Chennai | Reporting to the CTO

About Ekko

Ekko is an AI-powered healthcare product that turns clinical conversations into structured

documentation. Our pipeline transcribes audio, generates clinical notes, and extracts structured

data using LLMs. We ship a web app and a backend service used by clinicians who depend on

getting it right.

We're a small, early-stage team that cares a lot about craft, moves fast, and keeps the chai flowing.

The person in this role owns the part of Ekko that makes it magic: the AI itself.

The Role

We're hiring an AI Engineer to own and improve our LLM pipeline end to end. This is the core of the

product. A clinical conversation goes in, and accurate, structured documentation has to come out.

You'll work on every stage: transcription quality, note generation, and structured data extraction,

plus the prompts, models, and evaluation that hold it all together.

Accuracy is everything here. A hallucinated medication or a misattributed instruction is not an

acceptable bug in a clinical setting. Your job is to make our AI outputs reliably correct, measure that

rigorously, and keep raising the bar as we grow. You'll work closely with backend engineering

(where the pipeline runs), QA (on output validation), and product (on what "good" means clinically).

This is a hands-on, pragmatic role. We care about outputs that work in the real world, not novelty for

its own sake.

What You'll Do

Own the LLM pipeline

Build and improve the stages that turn audio into clinical notes and structured data

Engineer, test, and version prompts for accuracy, consistency, and safety

Select and tune models and providers (OpenAI and similar) for quality, latency, and cost

Get transcription and extraction right

Improve transcription accuracy across speech-to-text providers (AWS, Google, Azure)

Build robust extraction of structured fields from messy, real-world conversations

Design the pipeline to fail safely: handle ambiguity, gaps, and low-confidence cases gracefully

Measure quality rigorously

Build evaluation harnesses that compare AI output against ground-truth references

Define and track the metrics that matter clinically (accuracy, hallucination rate, omissions,

extraction correctness)

Use evals to drive real improvements, not vibes

Partner across the team

Work with backend engineering to ship pipeline changes reliably to production

Work with QA on validating AI output and catching regressions

Work with product to define what a "good" note looks like for clinicians

What You Bring

3+ years building with ML/AI, including hands-on work with LLMs in production or applied

settings

Strong Python, and comfort working within a production codebase

Real prompt engineering experience: you know how to make models reliable, not just

impressive in a demo

Experience evaluating AI output systematically (ground-truth comparison, semantic similarity,

structured validation, or similar)

A pragmatic, accuracy-first mindset and the rigor to measure what you ship

Comfort with startup life: ambiguity, ownership, and shipping over perfection

Nice to Have

Healthcare, clinical, or other high-stakes, accuracy-critical AI experience

Experience with speech-to-text / ASR systems and audio data

Familiarity with LLM evaluation tooling and techniques (RAGAS, LLM-as-judge, and similar)

Comfort working alongside a TypeScript/NestJS backend that hosts the pipeline

Familiarity with HIPAA, NABH, DPDP, or similar compliance frameworks

Serious foodies welcome; we take team lunch a little too seriously

Bonus if you can make an LLM behave with Rajinikanth-level confidence

What We Value

These are the values we hire for and live by:

People First, Always. Empowered, supported people build transformative solutions, so we

invest in our team.

Build with Empathy. We understand the pain points of clinicians and patients, and design

solutions that feel intuitive and add real value.

Pursue Practical Innovation. We innovate with purpose, not novelty, grounded in reliability,

simplicity, and outcomes.

Move with Humility and Urgency. We're ambitious about hard problems but stay grounded in

feedback. We ship fast, learn quickly, and adapt.

Think Long-Term, Act Today. We're building a platform for the future of healthcare, making

decisions with tomorrow in mind without losing momentum today.

Win Together. Great outcomes come from collaboration. Teammates, customers, or partners,

we listen, share, and grow as one.

Why Join

Real ownership. You own the AI that makes Ekko work, not a small slice of someone else's

model.

AI product depth. You'll solve accuracy problems most AI roles never face, in a setting where

correctness truly matters.

Direct impact. Small team, fast feedback, decisions made in days not quarters.

Good people. We take the work seriously and ourselves a lot less so.