Senior AI Engineer & Full Stack Developer with 10+ years of experience delivering AI-powered enterprise solutions across Latin America and North America.
• AI Engineering: Expert in building scalable multi-cloud AI systems, including RAG, agentic workflows, and LLM fine-tuning.
• Industry Experience: Proven track record delivering AI solutions in the mining and energy sectors.
• Product Delivery: Launched high-traffic platforms such as JournalAI.pro.
• Full-Stack Development: Skilled in Python, Node.js and React, with experience across AWS and Azure.
• AI Infrastructure: Hands-on experience architecting production-grade AI applications and deployment pipelines.
Builds enterprise RAG systems, optimising retrieval and contextual generation with LLMs.
Designs robust APIs that expose AI tools and integrate intelligent agents into enterprise environments.
Orchestrates autonomous agents with MCP across multiple models and specialised tools.
Delivers end-to-end ML pipelines, from data preparation through deployment and production monitoring.
Applies NLP and multimodal AI to business-critical enterprise use cases.
Advises enterprise clients in the mining and energy sectors on applied AI solutions.
Designs AI-driven workflows for large-scale mining operations, focused on process optimisation and automated reporting.
Architects multi-cloud AI solutions across AWS and Microsoft Azure, using Azure OpenAI Service and AWS SageMaker.
Builds enterprise OpenAI applications for conversational AI and document intelligence use cases.
Led JournalAI.pro from concept through launch, creating an AI-powered platform that automates high-quality SEO content generation.
Architected the platform’s AI workflows using LangChain and fine-tuned LLMs, tailored specifically to SEO-focused content generation.
Designed and deployed the supporting cloud infrastructure across AWS and Vercel, scaling the platform to 5,000+ active users across global markets.
Established analytics and A/B testing frameworks to measure user behaviour, improve engagement, and optimise overall platform performance.
Fine-tuned LLMs for domain-specific use cases, optimising performance for specialised business requirements.
Architected OpenAI API integrations for NLP, content generation, automation, and other AI-powered capabilities.
Led full-stack AI application delivery, translating complex requirements into scalable and reliable solutions.
Optimised AI model efficiency and accuracy, reducing computational costs and latency across production use cases.
Worked on the development and delivery of ML-enabled platforms for enterprise clients, contributing across backend development, cloud deployments, CI/CD processes, and scalable application architecture. He supported the integration of machine learning capabilities into business platforms while managing end-to-end software delivery workflows.
Founded and led a technology consulting and development agency focused on helping startups and tech companies build and scale digital products. He worked directly with clients on web and app development strategy, UI/UX design, and scalable application delivery, while also supporting product positioning and growth initiatives. In addition to managing client relationships and operations, he guided full product development workflows and helped companies integrate modern digital solutions aligned with their business goals.
Worked as COO overseeing the delivery and growth of digital marketing and web development services across multiple client accounts. He coordinated cross-functional teams across SEO, SEM, design, and development projects, while also helping align technical execution with business and marketing goals. His role included process optimization, client management, workflow implementation, and supporting the development of scalable digital solutions that improved operational efficiency and client retention.
Led the company’s digital transformation initiatives, modernizing internal operations through the implementation of ERP systems, infrastructure upgrades, and process automation. He worked closely with different business units to integrate new digital tools into daily operations, improve production and inventory workflows, and support company-wide technology adoption through training and operational improvements.
- Microsoft Certified: Azure AI Engineer Associate
- Microsoft Certified: Fabric Analytics Engineer Associate
- AWS Cloud Solutions Architect
- Microsoft AI & ML Engineering
Matias is a senior AI engineer and full stack developer with over a decade of experience delivering AI-powered enterprise solutions across Latin America and North America. He has a strong background in building scalable multi-cloud AI systems, including retrieval-augmented generation (RAG), agentic workflows, and large language model (LLM) fine-tuning. His recent roles at The Hackett Group and PwC Chile highlight his expertise in enterprise AI applications, particularly in mining and energy sectors.
He demonstrated deep technical knowledge and practical experience with advanced AI frameworks such as LangChain and LangGraph, as well as strong backend skills using Python, Node.js, and PostgreSQL. Matias also showed a clear understanding of system design trade-offs, API quota management, and observability practices, which are critical for production-grade AI systems. His communication was confident and clear, especially when discussing complex AI concepts and stakeholder management.
Matias’s approach to mentorship and leadership is adaptive and thoughtful, focusing on individual learning paths and practical guidance. He appears well suited to fast-paced, client-facing environments where technical depth and business acumen are equally valued, making him a strong candidate for senior AI engineering roles that require both hands-on development and strategic collaboration.
Deep experience with agentic AI frameworks. Matias demonstrated a thorough understanding of LangChain and LangGraph, showing how he leverages these tools to build autonomous agents and manage complex workflows. He is familiar with the Claude ecosystem and has transitioned to Anthropic models for their superior business capabilities and context handling.
Advanced RAG system design and optimization. He explained his use of fallback thresholds, tokenization parameters, and micro-algorithms to evaluate non-deterministic LLM outputs effectively. This reflects a mature approach to building reliable retrieval-augmented generation systems that maintain accuracy and performance over time.
Production readiness and error handling. Matias proactively anticipates API quotas and throttling issues, implementing jitter techniques, queues, and fallback models to ensure system stability under load. This practical mindset is essential for maintaining uptime and user experience in enterprise AI deployments.
He will thrive in a fast-paced agency environment.
Strong backend skills with Python and Node.js. Matias actively uses PostgreSQL and pgvector for vector search, correctly identifying best practices for multi-tenant architectures, including ID parametrization and middleware filtration to prevent SQL injection and data leakage.
Cloud and DevOps integration. He has hands-on experience deploying scalable AI applications across AWS and Azure, including monitoring and observability using tools like Phoenix and CloudWatch. This ensures robust production pipelines and cost management.
Security and observability focus. Matias prioritizes token security to prevent prompt injection attacks and implements detailed tracing and logging to maintain system health and facilitate debugging.
You handle the world with any software development. We don't have installability, honestly.
Adaptive mentorship style. Matias emphasises tailoring guidance to individual learning paths, recognising that each engineer has unique strengths and challenges. He shared examples of successfully mentoring junior engineers from bootcamps to full-time AI consulting roles.
Strong communication and stakeholder management. He is comfortable translating complex AI concepts for business stakeholders and managing expectations around project scope and changes, which is critical for leadership roles in AI initiatives.
Consulting mindset and coachability. Matias thrives in client-facing roles and fast-paced environments, showing a high degree of coachability and a desire to avoid bureaucracy while delivering value quickly.
I can adapt to any situation compared to most of the people.
Matias’s knowledge of some specific tools and frameworks could deepen further to enhance his already strong foundation.
- Supabase Row Level Security (RLS): His explanation was conceptual rather than detailed on exact policy syntax, suggesting room for more hands-on experience.
- Experience with emerging LLM frameworks: He has not worked with VLLM but is aware of it, indicating an opportunity to expand into newer inference engines.
These areas represent natural growth paths rather than gaps that would hinder immediate contribution.
Matias is a highly capable senior AI engineer with a rare combination of deep technical expertise, practical production experience, and strong communication skills. He is particularly well aligned to roles that require building and scaling complex AI systems using agentic frameworks and RAG architectures, especially in enterprise contexts.
His adaptive mentorship approach and consulting mindset make him a strong candidate for positions involving technical leadership and collaboration with diverse stakeholders. Matias would likely excel in fast-paced, client-facing environments where both hands-on development and strategic guidance are valued.
Matt is a highly experienced Senior AI Builder with deep expertise in agentic AI frameworks such as LangChain and LangGraph, and a strong command of the Claude ecosystem, particularly Anthropic models like Claude 3.5 Sonnet/Opus. He demonstrates advanced knowledge in Retrieval-Augmented Generation (RAG) systems, Python, and vector databases like PostgreSQL with pgvector, applying sophisticated techniques for AI model optimization and system reliability.
He actively manages AI system robustness by anticipating API quotas and throttling, implementing fallback models, and prioritizing observability and token security. Matt's system design skills include balancing latency and reliability in agent handoffs, and he is comfortable working in fast-paced, client-facing environments with a focus on business goals and cost management.
- Uses agentic AI frameworks LangChain and LangGraph extensively for building AI agents.
- Works heavily with the Claude ecosystem, especially Anthropic models Claude 3.5 Sonnet/Opus, preferring them over OpenAI for business-driven capabilities and context window size.
- Employs advanced RAG optimization techniques including fallback thresholds (P95, P70), tokenization adjustments (top_p, top_k), and micro-algorithms for non-deterministic LLM evaluation.
- Implements robust error handling for API quotas and throttling using jitter techniques, queues, and fallback models to maintain autonomous agent stability.
- Prioritizes observability through tracing, console logging, and tools like Phoenix and LangSmith to monitor AI system performance and token usage.
- Designs AI systems with trade-offs between fast runtime-buffered states (LangGraph) and slower, reliable database-backed states (MongoDB/Postgres) depending on latency requirements.
- Proactively anticipates API quota limits and throttling errors (e.g., 429/500) and implements jitter and queue mechanisms to prevent agent crashes.
- Uses fallback models to ensure continuous operation during high-traffic periods.
- Employs tracing and console logging extensively to monitor AI system behavior and token consumption.
- Utilizes Phoenix and LangSmith tools for enhanced observability of AI workflows.
- Balances agent handoffs between fast runtime-buffered states and slower database-backed states based on latency budgets.
- Applies multi-tenant architecture principles including ID parametrization and middleware filtration to prevent data leakage and SQL injection.
- Implements token security measures to prevent prompt injection attacks.
- Uses middleware filtration and UUIDs to safeguard multi-tenant database access and prevent data leakage.