Open to opportunities

Matheus Silva

@matheussilva1

Message

AI Engineer building production-grade conversational agents and scalable, low-latency data-driven AI systems.

Brazil

Message

What I'm looking for

I'm seeking roles building production-grade conversational AI and data systems end-to-end, emphasizing scalable low-latency performance, RAG and knowledge-graph explainability, Rust for performance-critical components, and collaborative, product-focused teams.

I am an AI engineer with three years of experience designing high-performance, user-centric AI agents and data pipelines for production. I focus on building reliable end-to-end AI products that serve thousands of daily interactions with low latency and strong UX.

I have led applied research and production work that produced peer-reviewed publications and a deployed conversational text-to-SQL agent used by hundreds of engineers at Petrobras. My work spans retrieval-augmented generation (RAG), dynamic knowledge-graph construction for explainability, session-aware memory, and the orchestration of systems with LangGraph and FastAPI.

I enjoy solving performance-critical problems and am diving deeper into Rust for workloads that need extreme efficiency. I seek roles where I can drive production-grade agent architectures, improve observability and robustness, and collaborate cross-functionally to deliver measurable business impact.

Experience

Work history, roles, and key accomplishments

Current

AI/Data Engineer

Current

BTG Pactual

May 2025 - Present (3 months)

Developed Spark pipelines processing 200M+ monthly records and integrated them into Airflow DAGs on EKS/EMR to compute revenue and commissions. Designed a general-purpose RAG-based AI agent for the Revenue & Commissions team backed by a domain-specific knowledge graph and MCP + LangGraph orchestration.

Spark Airflow Python LangGraph Model Context Protocol (MCP)Knowledge Graphs RAG

AI Researcher/Engineer

Tecgraf/PUC-Rio

May 2024 - May 2025 (1 year)

Lead author of a novel evaluation framework for database conversational agents and co-author of a production-grade LLM-based natural language interface, with both papers accepted at international conferences. Built and evaluated LLM-driven text-to-SQL systems using RAG, dynamic knowledge-graph context expansion, session-aware memory, LangGraph orchestration, and FastAPI serving.

LangGraph fastAPI Knowledge Graphs Python Text To SQL RAG

AI Engineer

COGG.AI

May 2022 - May 2024 (2 years)

Built a Slack-integrated RAG chatbot serving 300+ active users to centralize institutional Q&A and implemented an initial clinical decision support system for otorhinolaryngology using knowledge graphs. Implemented systems with LangGraph and FastAPI to enable live streaming responses and tool orchestration.

LangGraph fastAPI RAG Knowledge Graphs Python Slack Graph RAG

Education

Degrees, certifications, and relevant coursework

Pontifical Catholic University of Rio de Janeiro

Master of Science, Computer Science (Agent Engineering)

2024 - 2025

Activities and societies: Master's thesis work on dialogue dataset generation and evaluation; related code: https://github.com/MatheusOliveiraSilva/dialogue-dataset-generator

Completed a Master of Science in Computer Science focused on agent engineering, conducting thesis work on dialogue dataset generation and evaluation for conversational database agents.