CAIO JAHN

Summary

Computer Engineering student at Georgia Tech, supported by the Behring full-ride scholarship. Originally from Porto Alegre, Brazil. Focused on embedded systems, neurotechnology, and practical intelligence—tools that strengthen human capability instead of competing for attention. This page collects current projects and direction.

About

I grew up in Porto Alegre, Brazil, a city that taught me to value resourcefulness, community, and the art of making things work with what you have. From a young age, I was drawn to systems—not just how they look on the surface, but how they function at a fundamental, mechanical level. This curiosity wasn't just idle; it drove me to participate intensely in scientific Olympiads, where I learned the discipline of rigorous problem-solving, the importance of first-principles thinking, and the joy of intellectual competition. Those early experiences shaped my worldview: that complex problems are just systems waiting to be understood and optimized.

At Georgia Tech, as a Behring Scholar, I’ve been able to channel that drive into Computer Engineering, a field that perfectly bridges the physical and the abstract. I’m fascinated by the intersection of hardware and software—specifically, how we can build tools that extend human capability rather than distract from it. My academic journey has been defined by a search for "practical intelligence"—technologies that solve real, tangible problems in the physical world. Whether it's designing embedded sensors or architecting local AI systems, I am motivated by the challenge of building robust, efficient infrastructure.

My background is a mix of academic rigor and entrepreneurial spirit. I co-founded POCT, a peer-learning platform, which taught me the operational side of engineering—how to build things that people actually use, how to manage logistics, and how to lead a team towards a shared vision. This experience instilled in me a deep respect for the "human" side of engineering; the best code in the world doesn't matter if it doesn't solve a user's problem.

Outside of code and circuits, I find balance in lifting, playing piano, and exploring the structured complexity of modular synthesizers. I believe in deep work, deliberate practice, and building systems that last. I am an avid reader of philosophy and sci-fi, constantly looking for new frameworks to understand the future we are building.

Career Goals

My professional interests sit squarely at the intersection of hardware, software, and intelligent tooling. I am driven to build systems that remain reliable and performant even as complexity scales. I see myself gravitating towards the software-heavy side of ECE—building the "brains" and tools that make hardware useful. I want to work on technology that respects human attention and enhances our agency in the physical world.

• 2025–2026: I plan to immerse myself in research and internships focused on embedded systems and applied intelligence. I want to get my hands dirty with real-world constraints—power, memory, latency—and learn from senior engineers who have shipped critical infrastructure. I am particularly interested in roles that involve firmware development, RTOS, or edge AI implementation.
• 2026–2027: I will broaden my technical scope, exploring neurotechnology and advanced sensing. My goal is to understand how we can interface more directly with the world and with our own biology. I plan to take advanced coursework in neuromorphic computing and signal processing to build a theoretical foundation for future interfaces.
• After Graduation: I aim to join an early-stage team, likely in California, working on purposeful technology—tools that solve hard problems in physical reality, not just digital optimization. I want to be in an environment where the pace is fast, the problems are novel, and the impact is tangible. Companies working on robotics, brain-computer interfaces, or next-generation compute infrastructure are my primary targets.
• Long Term: I envision leading a team that builds durable, privacy-focused computing systems that improve daily life without demanding constant attention. I want to build the "calm technology" of the future—systems that work seamlessly in the background to support human flourishing. I hope to eventually found a company dedicated to these principles.

Resume

Download Resume (PDF).

• Microclimate analysis — built Arduino sensing stations (MQ-7, DHT22, and related modules) to compare air-quality patterns in school regions across Brazil and Vancouver, processed with Python and Pandas.
• Notion Anchieta platform — created a structured academic workspace used by roughly 1,000 students preparing for the Brazilian national exam.
• POCT peer-learning venture — co-founded and led operations for a student business recognized 1st in Porto Alegre, 3rd in Rio Grande do Sul, and 3rd nationally.
• Scientific Olympiad — medals in national competitions; ongoing mentorship.
• Behring Scholar — full-ride scholarship awarded for academic performance and leadership trajectory.

Projects

Discovery Project: Ghola

View Full Project Page →

Ghola is a local, privacy-first voice assistant inspired by the Dune universe. It runs entirely on my Mac using Whisper for speech recognition, Ollama for local LLM inference, and Piper for text-to-speech.

The system demonstrates a fully offline AI pipeline, capable of transcribing voice commands, generating intelligent responses, and synthesizing speech with low latency. It serves as a testbed for exploring local-first computing and systems integration.

Technical Implementation

The project is built in Python, orchestrating a complex event loop that manages audio streams and model inference. I utilized faster-whisper for optimized speech-to-text, ensuring that transcription happens in near real-time. The "brain" is an 8-billion parameter Llama 3 model running via Ollama, prompted to maintain a specific persona. Audio output is handled by Piper, a neural TTS engine that offers a significant performance advantage over cloud-based APIs.

Impact

Ghola proves that powerful AI assistants do not need to rely on cloud servers that harvest user data. It represents a step towards "sovereign computing," where users own both the hardware and the intelligence that powers their daily interactions.

Microclimate Project

I designed and built custom Arduino-based microclimate stations equipped with MQ-7, DHT22, and other environmental sensors. The goal was to capture granular air-quality data that broad regional sensors often miss.

I deployed these stations in school zones across Brazil and Vancouver to compare pollution exposure in different urban plannings. The data was processed using Python and Pandas, revealing significant disparities in local air quality that correlated with traffic patterns and urban density.

Technical Implementation

The hardware consisted of an Arduino Uno interfacing with analog and digital sensors. I wrote C++ firmware to handle sensor polling, data averaging, and serial communication. On the software side, I developed a Python data pipeline using Pandas and Matplotlib to ingest the raw CSV logs, clean the data, and generate comparative visualizations of CO2 levels, humidity, and temperature variances.

Impact

This project highlighted the "invisible inequality" of air quality in urban environments. The findings were presented to local school boards, raising awareness about the need for better ventilation and traffic management around educational institutions.

Notion Anchieta Platform

Recognizing a gap in how students organized their study materials, I built Notion Anchieta, a comprehensive academic workspace and peer-learning hub. It wasn't just a template; it was a structured platform used by approximately 1,000 students preparing for the rigorous Brazilian national exam.

The system provided concise summaries, progress tracking tools, and a curated database of learning resources. More importantly, it democratized access to high-quality study materials, allowing students to exchange knowledge and support each other's growth.

Product Design & UX

I utilized Notion's advanced database features—relations, rollups, and formulas—to create a "gamified" study experience. The interface was designed to reduce cognitive load, presenting students with only the most relevant tasks for the day while tracking their long-term progress through visual dashboards.

Impact

The platform was adopted by a significant portion of the student body, creating a shared digital commons. It demonstrated how digital tools, when designed with empathy for the user's workflow, can foster a community of collaborative learning and significantly reduce academic stress.

POCT

I co-founded POCT, a student-run business that designed and sold multi-use portable pockets to combat pickpocketing in Porto Alegre. We identified a critical safety issue in our community—theft in crowded areas—and developed a practical, physical solution that integrated seamlessly with daily wear.

As the operations and strategy lead, I managed the entire product lifecycle, from sourcing materials to coordinating sales channels. We sold 918 units, achieving the highest revenue of any student company in Rio Grande do Sul.

Operational Strategy

We implemented a direct-to-consumer sales model combined with strategic partnerships at local events. My role involved optimizing our supply chain to meet high demand and analyzing sales data to refine our pitch. We treated the venture not just as a school project, but as a real startup with P&L responsibilities.

Impact

Our success earned us 1st place in Porto Alegre, 3rd in the state, and 3rd nationally. Beyond the awards, POCT taught me the fundamentals of entrepreneurship: identifying a real problem, building a product that solves it, and executing a sales strategy to bring it to market.