Co-creating AI prototypes and agent workflows with federal clients. Curious about what we can build next.
[ — ] Résumé Updated version coming soon
My work spans AI prototypes, machine learning, and software tools. This corner of the internet is where those interests meet a long-running fascination with space.
Home base
IBM · Federal Client Engineering
Foundation
Penn State · B.S. Computer Science · Math minor
In this universe
Engineering projects, life beyond work, and a few detours.
I'm Lucas, an Associate AI Engineer at IBM in Federal Client Engineering. My background spans client AI prototypes, machine learning, and software tools.
Penn State · B.S. in Computer Science Minor in Mathematics
A few stops along the way
IBM / Associate AI Engineer
– Present
I co-create proofs of concept and demos with federal clients, showing how IBM technology can fit into the workflows they want to build. My work includes agent orchestration workflows and applications, along with solutions using watsonx.ai and watsonx.data.
The work starts with the client’s desired workflow and develops through hands-on prototyping together.
IBM / AI Engineer Co-op
–
Worked on client prototypes with watsonx.ai, watsonx Orchestrate, and open-source language models, plus reusable demos and workshop material.
HackPSU / Co-Executive Director
–
Led nine organizing teams across sponsorship, marketing, logistics, and more for Penn State's student-run hackathon.
Broadcom / Software Engineering Intern
–
Built internal tools for chip-register management and an API to connect them with existing workplace technologies. Worked with chip-design specialists to understand the problem behind the software.
Nittany AI Alliance / Machine Learning Engineer
–
Worked on blood-donor analysis for the American Red Cross using Python and Azure. Received the Advance Excellence Award.
Personal software, client tools, and an award-winning learning prototype. Explore the public projects below; internal work is described at a high level.
Personal project / Lunar mission design
LunarOS
.-""-.
/ o . \
| . O |
\ o . /
'-..-'
2-minute walkthrough, recorded against the running application.
A browser workspace for exploring the Moon with NASA data, screening candidate settlement sites, and designing and simulating lunar missions.
Real lunar data
LRO terrain and imagery, Diviner temperature, modeled solar visibility, and USGS geology are served by a Python backend in lunar coordinates. Every value shown carries its source, and missing data stays missing.
Settlement screening
Candidate sites are ranked on native terrain slope across the whole Moon, with each candidate’s analysis cells drawn directly on a Three.js globe.
Mission simulation
A deterministic energy model runs a mission hour by hour. The interface explains each interval’s power flow and plays rover routes back in time.
A terminal-inspired note inbox for capturing thoughts quickly and organizing them later, with categories, reminders, and a dedicated review queue.
Capture through interruptions
New notes are written to a local queue before syncing with Supabase. Pending captures remain available when the network drops and retry on reconnect.
Shared web and native core
A React web app and React Native / Expo client share note, authentication, and category logic, with storage adapters for each platform.
A deliberate review flow
Review groups notes by overdue reminders, explicit flags, unsorted captures, and older notes worth revisiting. Time-based rules and capped groups keep the queue focused.
An early childhood learning tool with AI-generated choose-your-own-adventure stories, weaving life lessons into the choices a child makes.
What I built
A demo connecting a language model and a diffusion model to an interactive story experience, using React and Flask.
Result
Our team pitched Kinderverse to industry experts and won a $2,500 prototype and MVP prize from Nittany AI.
Scope
A learning prototype demonstrated through a pitch. The repository is not available to share publicly.
React · JavaScript · Flask · Python · Git
Generative AI · Interactive storytelling · Team project
Broadcom / Internal tooling / 2024
Register Database System
+--+--+--+
|01|10|01|
+--+--+--+
A graphical tool for viewing and updating chip-register settings, built for collaboration across design teams at Broadcom.
I built the register-management software and an API to connect it with existing workplace technologies. The work brought together a MySQL data layer, a Python/Tkinter interface, and the realities of chip-design workflows.
Python · Tkinter · MySQL · SQL · Git
Internal Broadcom project · Source code and internal materials are not available publicly.
Nittany AI Alliance / 2024
Red Cross Donor Site Prediction
.---+---.
| / / |
+---.---+
Machine learning and data analysis with the American Red Cross to identify potential blood-donation site locations from donor patterns.
As part of the Nittany AI Alliance team, I developed Python models and worked in an Agile process. The project used scikit-learn, PyTorch, and Azure to explore a practical question about where donation sites could be most useful.
Python · scikit-learn · PyTorch · Azure · Git
Focus / turning donor data into evidence for location decisions.
Personal project / Live
ASCII Galaxy Portfolio
You are here.
This portfolio brings character art, animation, and ordinary web controls together: an ASCII rocket, a scroll-controlled journey through space, and small discoveries along the route.
JavaScript · Canvas 2D · CSS · Node.js
. * .
o ---- O ---- @
earth moon mars
\
* --(o)--
rings
\
* beyond
Rendering
Character-by-character rocket and planet art on a canvas, with HTML controls for the portfolio.
Navigation
Scroll drives the route. A separate clock animates the scenery, so pausing ambience leaves navigation usable.
Practical details
Responsive reading zones, keyboard access, reduced motion, and saved discoveries. A Node.js script assembles the preview.
Built with HTML, CSS, and plain JavaScript. The scenery is rendered in ASCII characters.
I collect baseball, hockey, basketball, and football cards, along with Pokémon and One Piece TCG. My collection spans the sports I follow and the worlds I enjoy.
I follow Philadelphia’s teams across baseball, hockey, basketball, and football. There’s almost always a season to keep up with.
03 / Side projects
.---.
--| * |--
--|:::|--
'---'
| |
Agentic AI & building
I’m especially interested in agentic AI and how to use it effectively. Outside work, I keep learning new AI topics, exploring new technology, and designing projects of my own.
04 / Further out
. *
.-===-.
-( (@) )-
'-===-'
* .
The space frontier
Right now, I’m drawn to space databases and the return to the Moon—especially the prospect of a lunar base and industry beyond Earth. That curiosity is part of what inspired this galaxy.