AI products
Generative AI services, retrieval workflows, and practical agent experiences shaped for real users and production-ready teams.
AI • Robotics • Developer tools
I build AI and robotics systems, then measure how they behave.
Applied AI Engineer
Robotics Developer
ML / Software Engineer
I’m a software developer based in Mar del Plata, Argentina. My work spans generative AI, evaluation, RAG, cloud systems, computer vision, robotics, and developer tooling.
Build reusable systems
Run real workflows
Measure latency and cost
Learn from failures
About me
I focus on systems that are useful, testable, and grounded in the realities of real-world deployment.
I’m a software developer with a strong bias toward practical execution. My work connects generative AI, robotics, and developer experience—from cloud-based AI prototypes to embedded control ideas and agentic tooling built for real engineering workflows.
Most of my public work focuses on reusable product patterns: Anthropic ecosystem starter kits, AWS-based AI experiences, and experimental interfaces for computer vision, robotics, and control systems designed to support real execution.
Where I build
I work at the intersection of intelligent systems, robotics prototyping, and the tooling that helps advanced workflows move from experimentation to dependable, scalable implementation.
Robotics prototypes
Embedded control ideas, sensor-driven experiments, and computer vision concepts that connect software to physical systems with purpose.
Developer tooling
Structured workflows for modern engineering teams, including APIs, MCP patterns, automation, and intelligent infrastructure that ships cleanly.
Tools I use
A practical stack built around moving ideas quickly while keeping the architecture clear, maintainable, and grounded in execution.
Languages
- Python
- JavaScript / TypeScript
- PHP
- SQL
AI / LLM & Cloud
- Anthropic Claude APIs
- Amazon Bedrock
- AWS AI services
- RAG and prompt workflows
Robotics & Embedded
- Arduino
- Control loops
- Computer vision demos
- Sensor-driven prototypes
Developer tools
- GitHub Actions
- Model Context Protocol
- Subagents and hooks
- CLI tooling
Selected work
I focus on projects where technical depth, practical implementation, and credible evidence converge—particularly in evaluation systems, cloud AI, and computer vision that support informed decision-making.
Deep dive
AI evaluation
ai-evaluation-platform
Why this matters: Shows a structured, repeatable way to evaluate LLM systems with reusable metrics, clear reporting, and a reproducible local workflow that supports stronger engineering decisions.
The repository is structured around dataset-driven benchmarking, provider abstraction, metrics, reporting, SQLite persistence, FastAPI access, and a Streamlit dashboard.
FastAPI
Streamlit
SQLite
pytest
View repository
AI engineering
ai-engineering-lab
Why this matters: Connects AI projects through a measurable workflow for building, evaluating, benchmarking, and improving systems, with clear separation between measured and illustrative results.
View repository MCP
mcp-advanced-starter
An advanced starter for Model Context Protocol concepts, including transport, roots, elicitation, and server-initiated sampling for richer agent workflows.
View repository Robotics perception
robot-vision-dashboard
Why this matters: Shows a perception pipeline rather than an isolated model demo: video input flows through browser inference, tracking, telemetry, and a live dashboard.
View repository E-commerce optimization
The-AI-s-Mercado-Libre-in-develop
Analyzes Mercado Libre product and advertising performance, surfaces profitability issues, and ranks products with actionable scale, optimize, or pause recommendations.
View repository Anthropic Academy
claude-api-starter
A practical Python starter kit for Claude API patterns such as prompts, conversation state, streaming, tool use, and agent loops.
View repositoryLet’s connect
If you’re hiring, building something adjacent, or looking for a developer who can work across AI, robotics, and practical software systems, I’d welcome the opportunity to connect and discuss where this experience can create value.
Email: lazarogvitolo@gmail.com