Machine Learning Engineer

About

We’re looking for a hands‑on Machine Learning Engineer who can build, deploy, and iterate on real‑world ML systems inside modern data platforms like Microsoft Fabric and BigQuery.

This is not a research role. You’ll work directly with messy operational data, build models that drive decisions, and integrate intelligent systems into production workflows.

You’ll also design retrieval‑augmented systems (RAG) that combine structured warehouse data with unstructured context to power real decision‑making—not just generate outputs.

You should be equally comfortable with:

• Data pipelines and warehouse architecture
• Applied machine learning
• Agentic AI systems that take action on data

Responsibilities

• Build and deploy ML models directly within data warehouse environments (Fabric, BigQuery, etc.)
• Design pipelines that transform raw operational data into usable features
• Identify patterns, anomalies, and optimization opportunities in large datasets
• Integrate ML outputs into downstream systems via APIs, workflows, and automation
• Develop agentic AI systems that interpret data, make decisions, and trigger actions
• Work with engineering and operations teams to productionize models (not just prototype them)
• Continuously improve model performance, reliability, and business impact
• Design and implement RAG pipelines combining warehouse data with unstructured sources (documents, logs, operational data)
• Build systems where LLMs retrieve, reason, and act on internal data

Requirements (Core Skills)

• Strong experience with machine learning in production environments
• Hands‑on experience with BigQuery ML, Microsoft Fabric, or similar data‑native ML tools
• Proficiency in Python
• Deep understanding of data modeling, feature engineering, and pipeline design
• Experience working with APIs, webhooks, and system integrations
• Agentic AI / Modern AI Stack
• Experience building or working with LLM‑powered systems
• Strong understanding of RAG architectures (chunking, embeddings, retrieval strategies, evaluation)
• Ability to combine structured (SQL/warehouse) and unstructured (vector/semantic) data
• Experience designing systems where models don’t just predict—but take action
• Data & Systems Thinking
• Comfortable working inside data warehouses as the primary compute layer
• Experience with ELT pipelines (dbt, Airbyte, or custom pipelines)
• Strong SQL skills (non‑negotiable)

Bonus Points

• Experience with real‑time or near real‑time decision systems
• Familiarity with workflow/orchestration tools (n8n, Airflow, Prefect)
• Experience in manufacturing, operations, or scheduling systems
• Exposure to vector databases, embeddings, or hybrid ML + LLM systems

What Success Looks Like

• You turn messy data into working systems quickly
• You ship production systems—not just notebooks
• You think in end‑to‑end systems, not isolated models
• You own problems from data ingestion → model → action

Who This Is NOT For

• Agencies of any kind
• Purely academic or research‑focused ML engineers
• People who only build models in notebooks and don’t deploy
• Engineers without strong data or SQL experience
• Candidates whose experience is limited to basic RAG demos or prompt engineering without real pipelines