Role & responsibilities

We are hiring an applied AI Engineer to build and harden the AI layer of an enterprise compliance platform on Azure OpenAI and Microsoft Power Platform. This is hands-on Azure OpenAI engineering: wiring and hardening LLM calls, enforcing structured output, building the vision path for scanned documents, and tuning document classification to a measured accuracy bar.

Please read this before applying: this is an applied LLM integration and prompt-engineering role. It does not involve training models from scratch, large-scale MLOps, or statistical modelling. We are looking for engineers who ship reliable, auditable AI features on Azure OpenAI, not research profiles.

Key Responsibilities

- Build and harden Azure OpenAI (GPT-5o text and vision) integrations: deployments, api-version, quota, and chat completions.
- Implement the vision path for scanned documents via the REST chat-completions endpoint, with all secrets held in Azure Key Vault.
- Engineer prompts for strict structured output: JSON-only contracts, schema validation, deterministic field extraction, and no invented facts.
- Build a lightweight evaluation harness to measure document-classification accuracy against a labelled dataset, tune the confidence threshold, and iterate in measured cycles to meet the accuracy target.
- Integrate AI calls from Power Automate flows and Azure Functions,



with error handling, retries, and a triage fallback for low-confidence results.
- Enforce data-handling discipline: PII masking before calls, and an audit record capturing model, version and confidence for every call.

Must-Have Skills

- Hands-on production experience with Azure OpenAI specifically (not only the public OpenAI API): deployments, api-version, quota, chat completions, GPT-5o text and vision.
- Multimodal / vision calls via the raw REST chat-completions endpoint, with an understanding that standard low-code connectors do not accept images, and keys held in Key Vault.
- Robust prompt engineering with structured / JSON output and schema enforcement.
- Experience building an evaluation or accuracy-measurement harness for an LLM classification or extraction task, including confidence thresholds and iterative tuning.
- Integration engineering from Power Automate flows or Azure Functions: HTTP actions, error handling, retries.
- Discipline around PII masking and audit logging.

Good-to-Have Skills

- Power Platform familiarity (Power Automate, Dataverse).
- Document-processing experience: OCR, splitting merged PDFs, vision on scanned pages.
- Prior exposure to legal, insurance, or another regulated domain.
- C# or JavaScript for Azure Functions.

📌 AI Engineer - Azure OpenAI ( LLM Integration) (Bengaluru)
🏢 Svavvashaa Technologies
📍 Bengaluru