Staff AI/ML engineer & creative technologist.
I’ve spent nearly a decade taking AI ideas from first prototype to working software: tools for the teams that run clinical trials, the platform that monitored the largest dam removal in history, and several generative art projects of my own, including an autonomous AI gallery and a mathematical print studio.
Explore my workCurrently a Staff AI/ML Engineer at Fortrea’s AI/ML Rapid Development Studio.
Professional work and independent projects.
AI tools for the teams that run clinical trials, taken from prototype to production. I’m a co-inventor on two published patent applications.
Our studio builds AI, AR/VR, and mobile applications.
Staff AI/ML Engineer
A studio that runs entirely on my own machine, turning still images into ambient video clips with versioned edits, masked effects, sound mixing, and export.
As Gybe’s first developer and later Head of Software, I built and ran the platform behind GybeMaps: water-quality monitoring for utilities, conservation groups, and the Klamath dam removal.
Full-stack work across an AI startup studio: clinical criteria review, graph data pipelines, live sensor charts, and editable Unity scenes.
Five fictional artists inhabit a wall-mounted museum, generating images and gallery notes locally on a Raspberry Pi.
A local mathematical-art studio for exploring equations, composing print collections, and planning an office wall.
Try a broader skill, such as Python, local inference, or leadership.
Select a tool to find related work or experience.
I build retrieval-based and agentic AI applications, integrate local vision and language models, and coordinate generation on a shared GPU. My ComfyUI work also includes a custom node using PyTorch tensors.
I build interfaces, services, and data workflows across professional and independent software projects.
I built daily satellite ingestion at terabyte scale, brought sensor observations into the same platform, and developed ETL and background-processing workflows across other products.
At Verdant, I worked on entity and relationship extraction, clinical concepts, connected rule editors, and live sensor charts. The work joined data modeling with interfaces people could inspect.
From language-driven Unity scenes to a wall-mounted art museum. At Gybe, I integrated field-sensor data and occasionally helped install and calibrate instruments.
I built StillRoom’s mask editor, layered sound preview, image compositor, and export pipeline. With FORM, I work from mathematical expressions through composition, print proofs, and hanging plans.
I trained a character LoRA for visual consistency and built a ComfyUI pipeline using ControlNet, OpenPose, and AnimateDiff to produce exercise demonstration videos. My broader generative-media work also includes IP-Adapter and reference-guided synthesis.
My engineering experience includes retrieval-based and agentic AI applications, Azure OpenAI, Cosmos DB, classical machine learning, web and mobile frameworks, and Linux environments.
At both Gybe and Fortrea, I’ve set technical standards and interviewed candidates to help build the team. I also build foundations that other projects reuse and help shape decisions beyond my own team, and at Gybe I led the roadmap and a distributed team.
I’ve mentored engineers at Gybe and Fortrea, along with interns and junior developers, through code review, pairing, onboarding, and regular 1:1s. I also help colleagues outside engineering learn to work with AI.
Across my career, I’ve worked on evaluation, observability, and MLOps/LLMOps, and I use Jira and Azure DevOps to coordinate planning and delivery.
Published application · Co-inventor
Using relevant guidance documents to help a language model suggest improvements to a draft.
US20260161882A1
First filed December 10, 2024 · Published June 11, 2026
Published application · Co-inventor
Extracting content from a captured image and sending it to a webpage in real time.
US20260141169A1
First filed November 11, 2024 · Published May 21, 2026
Proactively optimizing clinical trial operations and aligning resources with integrated disease prevalence insights
FluSync explores how disease surveillance and site-level data can guide clinical trial operations. It brings WHO, CDC, and ECDC data into comparative views and follow-up workflows. The paper describes a conceptual tool and simulation.
I contributed to this paper as part of the AI/ML Rapid Development Studio and am credited in its acknowledgments.
By Nicholas Tufillaro and Adam Belmonte. A report connecting remote-sensing water-quality maps with field measurements.
I contributed to the preparation of this report and am credited in its acknowledgments.
I take AI ideas, including agentic workflows, from concept to production applications in Fortrea’s AI/ML Rapid Development Studio, a small team with company-wide impact across commercial, clinical, marketing, and other functions. I also mentor engineers and help colleagues outside engineering learn to work with AI. Co-inventor on two published patent applications.
Read about the studioLed the software roadmap and a distributed engineering team across the US West and East Coasts and Portugal, mentoring its engineers while staying hands-on in the backend.
Joined as the first developer and built Gybe’s cloud infrastructure from scratch across multiple cloud providers: terabyte-scale daily satellite ingestion, machine learning, and pipelines for custom field sensors, serving customers around the world.
Full-stack work across more than 10 projects in an AI startup studio, including clinical NLP and criteria review, graph data pipelines, live sensor charts, and text-to-Unity scenes.
Explore the studio workI was always impressed by his ability to judge what level of sophistication is required for any given feature — he really internalized the golden 80/20 rule.
He joined our startup as the first developer and handled all aspects of our complex stack, finding and addressing weaknesses proactively from day one.
