Machine Learning Engineer – Remote

About the role

BairesDev is seeking a senior Machine Learning Engineer to design and deliver end‑to‑end AI solutions, including NLP, LLM, and agentic workflows. You will work remotely with a global team, turning complex business problems into reliable, production‑grade models that run on Databricks and AWS.

Key responsibilities

• Co‑define AI solutions with product, engineering and data stakeholders, translating business goals into measurable metrics.
• Build data pipelines and feature‑engineering workflows using Python, PySpark and Databricks.
• Develop, train, evaluate and maintain traditional ML models as well as NLP/LLM solutions such as Transformers, embeddings, RAG and structured outputs.
• Design agentic workflows that connect LLMs to tools, APIs, retrieval systems and business rules, ensuring safety and observability.
• Package, deploy, monitor and retrain models on Databricks/MLflow and AWS, exposing FastAPI services for real‑time, batch and asynchronous workloads.

Required profile

• 7+ years of experience delivering end‑to‑end ML, NLP, LLM and agentic AI solutions.
• Deep expertise in NLP, text intelligence and LLM techniques (tokenization, embeddings, semantic search, RAG, function calling, hallucination mitigation).
• Strong background in traditional ML models (classification, regression, ranking, forecasting, anomaly detection, explainability).
• Proven experience designing multi‑step agentic AI workflows with memory management and human‑in‑the‑loop review.
• Extensive MLOps experience with Databricks, MLflow and AWS, covering CI/CD, monitoring, drift detection and cost optimization.
• Advanced proficiency in English.

Required skills

• Python
• PySpark
• Databricks
• MLflow
• AWS
• FastAPI
• Transformers
• Large Language Models (LLMs)
• NLP
• RAG (Retrieval‑Augmented Generation)
• CI/CD pipelines
• MLOps and model monitoring

What we offer

• 100% remote work from anywhere.
• Competitive compensation in USD or local currency.
• Hardware and software setup provided.