I strive to understand the human condition through engineering and beyond.

I'm a mechatronics engineer and roboticist with a software + ML background. My focus is building physically-intelligent biomechatronic systems through robotics and materials innovations, mainly for cardiac health / internal medicine and orthopaedics / biomechanics.

I believe in treating structure, geometry, and topology as first-class objects to be uncovered, not averaged away by black-box models. This philosophy drives how I approach ML and computation in scientific, robotic, and biomedical settings.

🇮🇹 🇪🇸 🇨🇷 · Yale '28 · building humanist tech through engineering & research.

At the Harris Orthopaedics Laboratory (MGH / Harvard Med), developing the next generation of implant materials and devices to improve orthopedic-surgery outcomes, focused on reducing post-operative infection risk through mechanical devices that leverage PK/PD optimization for local antibiotic delivery in intra-articular (in-joint) conditions.

Lead engineer for the Dynamic Electromechanical Bioreactor (DEB) at the Yale ICBL, combining mechatronics and biochemistry to simulate exercise-induced cardiac stress in human-engineered heart tissues, studying how Arrhythmogenic / Desmoplakin Cardiomyopathy progresses in endurance athletes. Programmable electromechanical loading, real-time force feedback, electrophysiology.

Soft, variable-stiffness burrowing robots at The Faboratory, treating granular-media mechanics, structure, and geometry as first-class design variables rather than averaging them away. Presented at IEEE RoboSoft and the Princeton Robotics Symposium; co-author on Spatial-MemER, an open-source spatial-reasoning extension to robot policies built with Stanford's IRIS Lab.

• Common Ground: Standardizing Granular Media Characteristics for Burrowing Robots · 2026 IEEE Int. Conference on Soft Robotics (RoboSoft), pp. 1005–1010.
  _{C. L. Le, F. Fonseca Cagnazzo, N. Ramos, E. I. Figueroa, C. Creager, R. Kramer-Bottiglio.}
• Uncorrelated Track (UCT) Processing: Efficient Algorithmic Treatment and Benchmarking Best Practices for Space Domain Awareness · Advanced Maui Optical and Space Surveillance (AMOS) Technologies Conference, 2025, p. 84.
  _{F. Fonseca, O. Pinhasi, Z. Zitzewitz.}

hyperpolymath.com · hyperpolymath-v2: A personal life-OS for people who refuse to specialize. Areas, projects, tasks, habits, and calendar unified under a natural-language agent (JARVIS, on Claude Sonnet 4.6). Web + Tauri desktop + a custom ESP32 macropad. Open-source!

degreeint.com: DegreeIntelligence. Degree planning for Yale, used by ~1 in 6 Yale undergrads. Scrapes scattered major requirements into one place: live distributional + major tracking, a drag-and-drop 4-year simulator, shared plans with friends, and an AI advisor (CLEO). Founder. Free forever. (Yale Daily News →)

Spatial-MemER: Open-source spatial extension to MemER for hierarchical VLA policies. Augments visual memory with an egocentric spatial map (camera pose via DPVO + forward kinematics, rendered bird's-eye-view keyframes) so the high-level policy can reason over where, not just what. Built with Mark Music (Stanford '28); ~a few lines on top of MemER, running at 1 Hz.

iSpy: 🥈 2nd place, Google × Yale SOM Hackathon. A wearable AI vision system for the blind & visually-impaired. Clip a camera to your cap, ask "Hey Hellen…", and get real-time scene description. Gemini vision + ElevenLabs voice + a FastAPI feed server + a React Native app. Builds a knowledge graph of what you've seen.

PosturePoke: 🏅 5th place (Hardware), YHack 2026. A wearable that physically pokes you when you slouch. 6-DOF IMU tracks spine angle; an escalation ladder runs from a beep → library fart sounds → an AI voice coach → a servo-powered poker. Arduino + sensor fusion + a Next.js / Web Serial dashboard with no backend.

diush: Open-source mobile platform (1517 / Medici grantee): a link-sharing funnel that turns messy DM bidding into a clean, secure sale. React Native.

hyperpoly.me: A public, minimalist log of my daily language training, tracking reps across 7+ languages, treating each like a runner treats mileage.

• 🥈 2nd place worldwide, World Robot Olympiad (Senior Open, 🇨🇷): UpDrop, an autonomous drone-delivery system; historic finish for a Latin American nation, covered by 15+ outlets
• 🛰️ SDA TAP Lab Team & Research Lead: Uncorrelated Track (UCT) satellite-tracking algorithms with US Space Command (via YUAA)
• ⚗️ Summer Research Fellow: UCLouvain (IMMC/IMAP), multiscale reactor modeling for CO₂ methanation with green hydrogen
• 🧭 Software Engineer: CourseTable, Yale's most-used student site (millions of requests/semester); shipped the professor-insight page
• 🎙️ Frontend Systems & Architecture Engineer: Voqo AI (Sydney), B2B voice-AI platform
• 🤝 Founder & Head of Engineering, HSC-GPT: Australia's first localized AI education platform; acquired by Art of Smart Education
• 🌐 Co-Founder & CEO, Zyndicate: real-time collaboration platform, ~30k peak users, team of 12+ across 5 continents
• ▶️ Creator: YouTube: Filippo Studies

languages   TypeScript · Python · C / C++ · Go · Rust · Dart · Swift · MATLAB
ml / sci    PyTorch · TensorFlow · NumPy · VLA / embodied AI · geometric learning
web         React · Next.js · Node · GraphQL · Tailwind · Supabase / Postgres
mobile      Flutter · React Native · Tauri
robotics    mechatronics · control · sensor fusion · embedded (ESP32 / Arduino)
hardware    CAD · KiCad · prototyping · precision mechanical design