Marcos

I build and productionize AI/ML systems for startups and businesses.

Overview

Toronto, Canada

marcosai.dev

Work with me

Available for freelance and contract work

Start with a focused technical audit, extend an existing repository, or build a production-ready AI/ML system from the ground up.

Discuss your projectEmail meView projects

Services

  1. LLM Applications, RAG & Enterprise Search

    • Document Ingestion
    • Citations
    • Permission-Aware Retrieval
    • RAG Evaluation

  2. AI Agents & Workflow Automation

    • Human Review
    • API Integration
    • Document Processing
    • Business Rules

  3. Recommendation & Ranking Systems

    • Candidate Generation
    • Ranking
    • Cold Start
    • Drift Monitoring

Working together

I can work inside an existing repository, extend an MVP, improve an inherited AI system, or build on an appropriate open-source foundation. A full rewrite is not required.

Work can be hourly or fixed-scope. NDA, confidential data, and private repositories are supported. Typical delivery may include source code, documented APIs, environment setup, tests, deployment instructions, evaluation notes, and handoff material; exact deliverables depend on scope.

We'll define requirements and success criteria, agree on a technical plan and milestones, then evaluate, deploy where scoped, and document the handoff. Source-code delivery, documentation, and IP terms can be agreed in the contract.

GitHub contributions

Selected case studies
5
Service areas
6

Commercial work may be kept in private repositories for confidentiality; the public graph is an activity record, not a measure of project quality.

Now

I'm currently focused on grounded retrieval, practical evaluation, efficient model serving, and the reliability work that turns AI prototypes into maintainable production systems.

Discuss your project

Stack

Languages
AI Interfaces
Infrastructure & Data
LLM Tooling
ML & Deep Learning
MLOps

Projects(5)

  • Automated extraction, validation, and human review for business documents.

    • Built an ingestion and extraction pipeline on the open-source Docling foundation
    • Added stateful human-review orchestration and workflow checkpointing with LangGraph
    • Designed business rule engines for 3-way matching, duplicate prevention, and mock ERP connectors
    Problem
    Invoices and purchase orders arrive in inconsistent formats and need extraction, review, validation, and delivery into a business workflow.
    My contribution
    Built the application layer around open-source Docling and LangGraph: ingestion, extraction workflows, human-review checkpoints, business-rule validation, duplicate checks, and mock ERP connectors.
    Validation / outcome
    Demonstrates a reviewable document-to-ERP workflow with checkpointed orchestration, rule-based validation, duplicate handling, and clear paths for human intervention.
    • Python
    • LangGraph
    • vLLM
    • Docling
    • Docker

    Need a document workflow adapted to your forms and business rules? Discuss your project.

View all projects