AI Engineer

Core responsibilities

• Scope LLM features with a clear task, a success metric and a fallback
• Build retrieval pipelines: chunking, embeddings, vector search and reranking
• Write and version prompts and tool definitions as code
• Build eval sets and run them on every change to prompts or models
• Add guardrails for safety, PII and hallucination, with human review where needed
• Track token cost and latency, and route between models to control both
• Ship features behind flags and measure them with real users

Skills that separate strong AI engineers

Skill	What good looks like
Problem framing	Defines the task, the metric and a non-LLM baseline before building
Retrieval	Tunes chunking and reranking with measured recall
Evaluation	Keeps a golden set and blocks releases that regress it
Cost and latency	Knows the per-request cost and can cut it with caching or a smaller model
Production engineering	Writes tested, observable services with retries and timeouts
Communication	Explains model limits to product and support teams in plain words

Interview questions and how Exordiom vets

1. Walk me through an LLM feature you shipped. How did you measure quality before launch?
2. Retrieval quality dropped after a document update. What would you check first?
3. How would you cut the cost of a feature that calls a large model on every request?
4. When would you fine-tune instead of improving the prompt or retrieval?
5. How do you stop a model from leaking customer data across tenants?
6. Explain a model failure to a non-technical stakeholder in two sentences.

Red flags: no evals, quality judged by reading a few outputs, and no idea what a request costs.

Exordiom recruits employed engineers, never a bench. Each candidate completes an AI interview built from your spec, then human rounds that score character, acumen, grit and experience. After the start, Hypercare check-ins run at week 1, week 2 and every four weeks from week 6, with you and your hire separately.

India vs Philippines for AI engineers

India is Exordiom's main market for AI and engineering roles, and companies such as Rubrik, Databricks, Anthropic and OpenAI list applied AI roles in Bengaluru, Mumbai and Delhi. In the Philippines, Netskope is hiring AI agent developers in Taguig and Salmon is hiring ML engineers in Metro Manila. In both markets your engineer works a full shift on US hours. For more on the India market, read India's AI engineering talent pool.