Building the machines that move people safely

Bucky's Blast Off

A working model of a pneumatic launch coaster, built on a 31-member team where I led the subsystem leads and personally sat on the launch team doing the launch design and analysis. The mechanism is a compressed-air launch driving a launch dog into the train, the same architecture as MAXX Force at Six Flags Great America. Rather than a full circuit, the layout is a single vertical spike, a choice we made deliberately: a spike shows kids the physics of a launch, kinetic energy trading for height and rolling back, far more clearly than a sprawling layout would.

The launch was engineered to a hard target. I worked the kinematics to size the launch so the train exits at exactly the velocity needed to reach the top of the 26″ spike and roll back cleanly, no more and no less. That meant calculating launch acceleration, exit velocity, and spike clearance, then reconciling them against the losses in the real system. The biggest of those losses was the surprise: track friction during fabrication came in far higher than predicted and was killing our exit speed. The fix was unglamorous and effective. Graphite powder on the track dropped the friction enough to hit the target, and it stuck.

The train itself is a themed Bucky rocket ship, 4″ long and about 2.5″ wide, carrying the club's professor as its rider. It was demonstrated at Engineering Expo as part of the booth that won Best in Show.

The number I point to is reliability: the model has run over 5,000 cycles across two years of demonstration with zero maintenance. For a student-built mechanism with a pneumatic launch and moving train, that is the real proof it was engineered, not just assembled.

ASTM F24.80 Harmonization — Historical Change Record

Ongoing work I am doing on a two-person team through the Next-Gen ASTM group, contributing to F24.80, the harmonization standard for amusement rides and devices. The goal is to build a historical record and change log that captures everything that changed between versions of the standard, and, more importantly, documents why each change was made. Standards evolve version to version, but the reasoning behind a given revision often lives only in the memory of the people who were in the room. Once that context is lost, future committees end up relitigating decisions that were already settled for good reasons. This record is meant to prevent that, giving future standard development a clear line of sight into the intent behind the language it inherits.

The part that makes this real work rather than a paperwork exercise is how we are sourcing the “why.” Rather than guessing at intent from the redlines, we are conducting technical interviews with the people who actually made the changes, and pairing what they tell us with the version history so each documented change is tied to the reasoning that drove it.

The project is in progress. What it represents is unusual for a student: direct involvement inside a working ASTM subcommittee, alongside the industry professionals who write and maintain the safety standards the amusement industry runs on, doing the kind of technical documentation and institutional-memory work that shapes how those standards develop going forward.

ASTM F2291 Class-4 Suspended Ride Vehicle

A tabletop working model of an inverted (“suspended”) coaster vehicle, built for Engineering Expo to teach local children how these ride mechanisms work and, more importantly, how they keep riders safe. I led a 5-member team on it. The model is a full ride vehicle plus a track segment and the supports it demonstrates on, presented the way manufacturers show off vehicles at IAAPA rather than as an isolated part on a table. It runs 24″ long and carries a 6″ scale patron.

The centerpiece is a dual-redundant, ratcheting over-the-shoulder restraint that is compliant with ASTM F2291 Class-4 both mechanically and electrically, so no single failure in either domain can release the rider. This restraint was the predecessor to the one I later designed for my robotic arm ride model, and it is where the approach was first worked out. The design is backed by real engineering: restraint load cases, structural analysis of the suspension carrying the vehicle under the track, and material and fabrication choices sized to the loads rather than guessed. The vehicle, track, and supports were produced with 3D printing, laser cutting, and CNC machining.

The part I am proudest of as team lead was not any single component. It was getting four other people to actually open the ASTM standard, read it, understand what Class-4 requires, and implement it correctly in hardware. Standards work is where most student projects wave their hands, and turning a dense specification into a fabricated, demonstrable mechanism that a team of five understood end to end is the real result.

The finished model was fabricated and demonstrated at Engineering Expo as part of the booth that won Best in Show.

36 sq ft Dark Ride

A fully realized dark ride built into a 36 sq ft footprint, with a 40 ft ride path, six show scenes, seven servo-driven animatronics, a themed show set, and a POV camera fixed to the vehicle so the whole thing can be experienced from the guest's seat as if it were full scale. I led the project as Operations Officer on a 40-member team, which meant I ran the subsystem team leads: owning the technical design across subsystems, keeping the track, vehicle, show, and controls teams integrated and unblocked, and steering the engineering and analysis that tied it all together.

The ride is a tracked, continuously moving system, and it tells a story. The vehicle is a submarine descending deeper and deeper into the ocean: it gets startled by a shark, is guided onward by bioluminescent fish, discovers a wrecked ship, and finishes with a conducting octopus. Each of the seven animatronics runs on its own Arduino Nano, all coordinated by a master ESP32 that sequences the animatronics against the vehicle's motion so scenes trigger on cue as the sub arrives.

Beyond the leadership role, I did hands-on engineering: I designed the ride enclosure and the cable routing for the whole model, and I led the track and ride vehicle team closely enough to be in the mechanical design, not just managing it.

The model was demonstrated at Engineering Expo as part of the booth that won Best in Show. This year it is being upgraded from the single continuously moving vehicle to a three-vehicle system with block zones and timed dispatches, along with additional animatronics, moving the model closer to how a real dark ride manages throughput and spacing.

Additive vs. Subtractive Manufacturing for Aircraft Engine Components

A technical report evaluating whether additive or subtractive manufacturing is the better fit for producing aircraft engine components, written to persuade a mixed audience of engineers and non-technical stakeholders.

I built the comparison around a weighted scoring of the criteria that actually drive a production decision: material efficiency, lead time, part strength, proven failure history, and scale economics. When the numbers came back, the two methods landed in a near tie. That result was more useful than a clean winner would have been, because it meant the real answer was not "which method is better" but "which method is better for which component." I owned the comparative analysis, the recommendation, and the conclusions and next-steps sections of the report.

My recommendation was scoped rather than universal. Additive manufacturing makes sense for components with complex internal geometries and lower production volumes, like fuel nozzles or parts with internal cooling routes that subtractive methods physically cannot produce. Those are exactly the cases where additive's lead-time and material-efficiency gains matter most, and where its weaker part-strength score and risk of layer separation matter least, since the parts run in small batches and can be inspected individually. Subtractive manufacturing should stay the standard for high-volume, high-risk parts like housings and brackets, where proven strength records and scale efficiency outweigh any material savings.

The recommendation only holds up if the risks are handled directly, so I paired it with three mitigations. Certification should roll out in phases, starting with non-critical components, so failure data can accumulate under low-stakes conditions before additive parts are trusted anywhere flight-critical. Z-pinning and finish machining should become standard practice for additive components, since both measurably close the fatigue gap between additive and wrought material. Adoption should come with workforce retraining, enabling machinists with deep subtractive experience get redirected onto additive lines rather than displaced. I also worked through the ethical dimension of that last point, since a manufacturing shift like this has real consequences for the people running the floor.

The report demonstrates technical writing for a split audience, analytical rigor in reaching a defensible recommendation from an ambiguous result, and working domain knowledge of aerospace manufacturing.

Attend.

A B2B Outlook add-in that plugs into Microsoft 365 and schedules mass meetings automatically, on a graph-theory optimizer I designed.

Attend is a startup I founded, funded, designed, and built myself. It came out of a year of professional engineering management work, where I watched a huge amount of time disappear into meetings: meetings that ran too long for what they accomplished, meetings too short to matter, meetings with the wrong people in the room, and, worst of all, the endless back-and-forth of just getting them scheduled. Coordinating a large meeting today means manually juggling everyone's availability, priority, and room constraints by hand, and no existing product actually solves that. They schedule what you tell them to. None of them find the optimal time and routing on their own.

So I built the thing that should exist. Attend is a B2B Outlook add-in that plugs into Microsoft 365 and schedules mass meetings automatically. At its core is a graph-theory-based optimizer I designed: it first guarantees every required attendee can actually make it, then maximizes how many optional attendees can join, then optimizes room selection by capacity and travel distance. It pulls live availability through the Microsoft Graph API, intersects everyone's free time, and returns the genuinely best slot rather than the first one that technically works.

The part I am most proud of is that I built all of it solo, and most of it was territory I had never touched. Going in, I knew basic Python and nothing else. To ship this I taught myself graph theory to build the optimizer, and the entire Azure hosting and backend world to run it: an async FastAPI backend, a multi-tenant PostgreSQL database with proper tenant isolation, secure Microsoft Entra authentication, and cloud hosting and storage. I funded the whole thing out of my own student accounts.

Attend is in pilot use now. More than the product itself, it is the clearest evidence of how I work: I find a real problem, I decide it is worth solving, and I go learn whatever I have to in order to solve it well.

5 DOF Robotic Arm Attraction Model

A solo-designed, desktop-scale working model of a “KUKA coaster” — the robotic-arm attraction type where a passenger capsule rides the end of a multi-axis industrial arm through a programmed motion profile. The scale patron is a 6″ Bucky Badger figure seated in a fully ASTM F2291 Class-4 compliant restraint, and the platform is modular, so one arm can run different ride profiles rather than being locked to a single motion.

The arm has five joints, giving it five degrees of freedom — one rotational axis per joint. Every joint is driven through a cycloidal reduction I designed from scratch: a 20:1 stage with 21 fixed ring pins acting against a 20-lobe disc, run as a dual-disc pair set 180° apart so the eccentric masses balance each other. I generated the disc profile mathematically and imported it into SolidWorks equation driven curves, then packaged it around a 2.43″ pitch circle with 0.04″ of eccentricity. I chose cycloidal over harmonic drives for the two reasons that matter on a ride: near-zero backlash for repeatable motion, and tolerance for the shock loads a start-stop ride profile throws at the joints.

Actuation is a NEMA 17 stepper per joint driven by TMC2209 drivers, with motor current set in software over UART rather than by trimpot, all coordinated by an ESP32. A GT2 timing-belt stage sits between each motor and its cycloidal input, which offsets the motor axis for tighter packaging and leaves room to add reduction if a profile needs it.

Because the ride profile I targeted put the model in Class-4 territory, the restraint had to meet that bar in full. I designed a dual-redundant, ratcheting over-the-shoulder lap bar that is compliant both mechanically and electrically, so a single failure in either domain does not release the patron.

The design is backed by real analysis, not just CAD: torque and load calculations at the joints, inverse kinematics for the arm, FEA on the loaded members, and a full dynamic-systems analysis of the ride arm through its motion.

Full CAD assembly is complete, the cycloidal drive is built, and the arm segments are in fabrication now.

Folding Projector Robot

A design project from my geometric modeling final, where the brief was to design a fictional piece of technology and take it from concept through full CAD. I designed a WALL-E-style projector robot built around one idea: it collapses down to the size of a water bottle for easy carrying, then unfolds — legs and a head deploy out of the packed body — to stand itself up and frame a projected image squarely against a wall.

The concept is backed by a real mechanism, not just a nice shape. The folding and deployment are worked out underneath the design, so the legs and head actually have a path from stowed to standing rather than magically rearranging. I ran a center-of-gravity check to confirm the deployed robot stands stably on its own legs, which is the constraint that makes or breaks a fold-out form like this.

I owned the design from paper to model: concept sketches, geometric dimensioning, orthographic drawings, and the mechanism layout, then led a 4-person team to take it into CAD. Leading the CAD effort meant setting the assembly structure, dividing the components across the team, and keeping everyone's parts consistent so the pieces actually fit together in the final assembly.

The project stayed a design study rather than a build, but it demonstrates concept-to-CAD design leadership: generating an original mechanism on paper and driving a team to realize it as a coherent, buildable model.