Postúlate en Kit Empleo: kitempleo.com.co/empleo/1asiwa

Medellín, Antioquia,Bogotá, Capital District,Bucaramanga, Santander,Cali, Valle del Cauca,Barranquilla

About project

**Provectus is an AWS Premier Consulting Partner and AI consultancy featured in Forrester's AI Technical Services Landscape, with 15+ years of experience and 400+ engineers. We build production AI for general enterprises in partnership with Anthropic, Cohere, and AWS.**:
**Role Purpose**:
As a Mid-Level ML Engineer at Provectus, you will work with increasing independence to design, implement, and deploy production-grade ML solutions for our clients. You sit at the bridge between learning and leading: you no longer require task-by-task guidance, yet you continue to grow toward senior technical ownership. A defining characteristic of this role is proficiency in AI-assisted development. You will leverage AI coding tools, contribute to agentic engineering initiatives, and actively shape Provectus's internal AI toolkit. You will also mentor junior engineers and contribute meaningfully to technical design decisions.

Core Responsibilities:
**Technical Delivery (55%)**:

- Design and implement ML pipelines from experimentation to production with limited supervision
- Build, evaluate, and optimize models across supervised, unsupervised, and generative AI tasks
- Develop and maintain production-grade Python code: modular, tested, and well-documented
- Set up reproducible experimentation environments and maintain experiment pipelines
- Deploy and monitor ML models in production, ensuring stability and performance
- Leverage AI-assisted development tools to increase velocity and code quality on all tasks

**Agentic Engineering & AI-Assisted Development (20%)**:

- Claude Ecosystem Integration: practical use of Claude Code or the Claude Agent SDK to deliver high-quality greenfield customer engagements




- Transform existing brownfield projects into AI-friendly setups
- Active usage of the Provectus AI toolkit in daily workflows
- Internal Contributions: contribute back to the Provectus AI toolkit, developing specific agents, building MCP servers, submitting bug fixes, adding features, or improving documentation
- Agent Frameworks: hands-on experience with Amazon Bedrock AgentCore, Strands, CrewAI, or equivalent orchestration frameworks for building tool-using and multi-step agentic systems
- MCP Integration: understanding of Model Context Protocol and ability to integrate or build MCP servers for client or internal use
- Stay current with emerging AI coding tools and agentic frameworks, sharing relevant findings with the team

**Collaboration and Contribution (15%)**:

- Mentor and support junior ML engineers on technical tasks, code quality, and best practices
- Conduct meaningful code reviews with constructive, actionable feedback
- Collaborate with cross-functional teams: DevOps, Data Engineering, Solutions Architects
- Share knowledge through documentation, presentations, and internal workshops on AI tooling

**Innovation and Growth (10%)**:

- Stay current with ML research and emerging frameworks, especially in GenAI and agentic AI
- Propose improvements to existing solutions, pipelines, and team processes
- Contribute to the development of reusable ML accelerators and internal quick-starts




- Participate in technical design discussions and architectural trade-off conversations

Technical Requirements:

- **Machine Learning Core**:

- Strong grasp of supervised and unsupervised ML: algorithms, evaluation, and real-world trade-offs
- Practical experience with classification, regression, and feature engineering in production or near-production contexts
- Hands-on experience with deep learning: CNNs, RNNs, Transformers training and fine-tuning
- Solid understanding of model evaluation, bias-variance trade-offs, and validation strategies
- Experience with at least one ML domain in depth: NLP, Computer Vision, Recommendation, or Time Series
- **LLMs and Generative AI**:

- Hands-on experience designing and implementing RAG systems (chunking, embedding, retrieval, generation)
- Working knowledge of vector databases (OpenSearch, Pinecone, Chroma, FAISS) and embedding models
- Understanding of prompt engineering, chain-of-thought reasoning, and LLM evaluation techniques
- Awareness of Amazon Bedrock capabilities: model invocation, Knowledge Bases, and Agent capabilities
- **Agentic Engineering & AI-Assisted Development**:

- AI-Assisted Development: demonstrated proficiency with AI coding tools (Claude Code, Cursor, GitHub Copilot, or similar) not just autocomplete, but strategic use for generation, refactoring, debugging, and documentation
- Agent Frameworks: hands-on experience with Amazon Bedrock AgentCore, Strands, CrewAI, or similar orchestration frameworks; ability to build stateful, tool-using agents
- MCP Integration: working understanding of Model Context Protocol; ability to consume or contribute to MCP servers for internal or client-facing integrations
- Tool Use & Function Calling: practical exper

Postúlate en Kit Empleo: kitempleo.com.co/empleo/1asiwa

📌 Middle Ai/ Ml Engineer (Medellín)
🏢 Provectus IT
📍 Medellín