AI Engineer

Position Purpose:

A proficient AI Engineer will join our IT team, focusing on developing and enhancing AI Systems Engineer, you'll play a pivotal role in transitioning to an AI-driven company. Your work will encompass designing, developing, and ship production AI solutions across ML models and LLM systems—including AI agents, RAG, Agentic AI, and Agentic RAT (Agentic RAG / Retrieval-Augmented Tooling)—using Azure, OpenAI/Azure OpenAI, and Google Gemini.

 

Roles and responsibilities:

AI Agents + Agentic AI (Hands-on)

• Build tool-using agents that execute multi-step tasks: planning, tool calling, verification, retries/fallbacks, and audit logs.
• Implement agent orchestration (graph/state machine patterns), deterministic controls, and human-in-the-loop escalation.
  
  

RAG + Agentic RAT (Agentic RAG / Retrieval-Augmented Tooling)

• Build RAG pipelines end-to-end: ingestion, chunking, embeddings, vector/hybrid retrieval, reranking, citations, grounded responses.
• Implement Agentic RAG: retrieval and tool-use loops (“retrieve, reason, tool-call, verify, respond”) with confidence scoring. Tune retrieval quality: metadata filters, hybrid search, prompt grounding, evaluation datasets, and regression tests.

Data Science + Machine Learning (Hands-on)

• Own end-to-end ML: problem framing, EDA, feature engineering, training, validation → deployment , monitoring.
• Build ML models (classification/regression/ranking/forecasting) using scikit-learn and/or PyTorch/TensorFlow.
• Apply rigorous evaluation: cross-validation, leakage prevention, bias checks, calibration, thresholding, lift/uplift analysis.
• Create production-grade feature pipelines (batch + real-time where needed) and ensure reproducibility.
  
  

ML Deployment + MLOps (Hands-on)

• Deploy ML models as APIs/batch jobs (FastAPI/Azure Functions/containers) with performance and reliability.
• Implement MLOps: CI/CD for training + deployment, experiment tracking (MLflow or equivalent), model registry/versioning, rollback.
• Production monitoring: model drift, data quality checks, performance degradation alerts, latency/cost monitoring.
• Write runbooks, on-call-friendly dashboards, and incident playbooks for model failures.
  
  

Cloud + Model Providers

• Deploy on Azure: Blob/ADLS, Key Vault, Azure AI Search (vector/hybrid), App Service/AKS/Functions, App Insights.
• Use OpenAI/Azure OpenAI and Google Gemini with provider abstraction, prompt/version governance, and rate-limit handling.
  
  

Physical Demands:

Not Applicable.