Marcos
I help Canadian teams get ML models out of notebooks and into production.
Overview
Work with me
Available for work
Book a 20-minute fit call — we'll scope the problem, the timeline, and whether I'm the right person for it.
Services
Machine Learning Systems
- Custom ML Models
- Predictive Modeling
- Recommendation Systems
- ML Pipelines
AI Infrastructure & MLOps
- MLOps / LLMOps
- Model Deployment
- ML CI/CD
- Model Monitoring
- Cloud & GPU Infrastructure
LLM & Generative AI
- RAG & GraphRAG
- LLM Fine-Tuning
- LLM Evaluation
- AI Agents
- Private LLMs
References
References from recent clients are available on request — ask for them on our first call and I'll put you in touch directly.
Social links
GitHub contributions
- Years building
- 7+
- Projects shipped
- 5
- Awards
- 4
- Focus
- ML
Now
I'm currently helping teams build, deploy, optimize, and secure AI systems — from custom models and RAG applications to computer vision, GPU infrastructure, and workflow automation.
Start a projectStack
Languages
ML & Deep Learning
Experience
Northlane AI
- Location
- Toronto, Canada
- Location type
- (Remote)
- Employment status
- Current
- Lead the design of retrieval-augmented assistants over private enterprise documents, including chunking, hybrid search, reranking, and grounded answer evaluation.
- Own the production loop: training and fine-tuning pipelines, offline and online evaluation, deployment, monitoring, and drift alerts.
- Cut median LLM inference cost per request by routing between a fine-tuned small model and a frontier model based on task difficulty.
- Harden deployed assistants against prompt injection and data leakage with input/output filtering and red-team test suites.
- Python
- PyTorch
- LLM Fine-Tuning
- RAG
- Vector Databases
- Evaluation Harnesses
- LLM Security
- Kubernetes
Meridian Analytics
- Location
- Vancouver, Canada
- Location type
- (Hybrid)
- Python
- scikit-learn
- XGBoost
- Feature Stores
- Airflow
- MLflow
- AWS SageMaker
- Model Monitoring
- Python
- Docker
- MLOps
- CI/CD
- GPU Optimization
Vantage Vision Systems
- Location
- Calgary, Canada
- Location type
- (On-site)
- Python
- PyTorch
- OpenCV
- YOLO
- ONNX Runtime
- Edge Inference
- OCR
- Python
- pandas
- SQL
- Time Series
- Statistics
Projects(5)
Automated pipeline converting unstructured documents into actionable ERP data.
- Built a document ingestion pipeline with baseline and fine-tuned extraction models using Docling
- Developed a stateful orchestration layer via LangGraph for human-in-the-loop review and workflow checkpointing
- Designed business rule engines for 3-way matching, duplicate prevention, and mock ERP connectors
- Problem
- Processing unstructured documents like invoices and purchase orders required manual labor, leading to errors, financial losses, inventory discrepancies, and delayed vendor payments.
- Role
- Designed and built the end-to-end pipeline, covering document ingestion, model extraction, human review orchestration, business-rule validation, and ERP integration.
- Result
- Reached 98.2% total extraction accuracy and established a reliable automated document processing system capable of blocking duplicate attempts with a 99.9% successful completion rate.
- Python
- LangGraph
- vLLM
- Docling
- Docker