Machine Learning Engineer

ML Engineer | Wyra AI Labs
Bangalore · Full-time
Wyra builds NexuS OS, a decision platform for mid-market industrial and healthcare enterprises.
We read fragmented ERP, HIS and operational systems, understand what the business means, reason over that state, and write decisions back into the systems where work happens. Most of this isn't LLM work. Our agents commit real transactions. The hard part is not generating an answer. It is knowing whether the answer is correct enough to act on.
What you'll build
You'll build platform primitives, not per-customer fixes:
Entity resolution across ERP, HIS and operational systems
Ontology induction from raw schemas, tables and PDFs
Leakage and anomaly detection
Causal attribution
Forecasting on sparse, dirty histories
Agents with real state, retries and failure handling
The eval harness that gates every write-back
What we're looking for
2+ years shipping classical ML in production. Models you owned end to end, strong SQL, time series and statistical inference. Ranking, fraud, risk and demand forecasting all count.
The instinct to tell which part of a customer problem is specific to that customer and which part belongs in the product.
Judgement about when not to model.
Strong production engineering fundamentals. You should be comfortable with state, idempotency, retries, concurrency and failure recovery.
Patience to implement research when there is no usable reference code.
Comfort with small, sparse, undocumented datasets and modest on-prem compute.
We care less about how many ML frameworks you know and more about whether you can take a messy problem, figure out what is actually going on, and build something we can trust.
Apply
Email hello@wyralabs.ai with something that shows how you think. A repo, a write-up, a model you shipped, or a paper you implemented. Tell us what worked, what broke, and what you learned.

Work arrangement
No