JJ

JJ / EXIBIST

Robotics / AI / Systems

Engineer / StudentASL Robotics / FRC Team 1884

JJ / EXIBIST

Building robotics, AI, simulation, and software systems with an architect’s eye for the code, tooling, and operator loop.

I work across robot software, vision tooling, ML research, simulation architecture, and interface design. The thread through all of it is system thinking: how the codebase is shaped, how people use it, and whether it still holds up when the constraints get real.

Operating range

robotics softwareAI / MLsimulationvisionsystems designtechnical UI

Current signal

Public team software work across FRC 1884 repositories, plus independent systems and ML projects built outside the season cycle.

Strongest through-line: architecture choices that make hard systems easier to run, reason about, and extend.

Focus areas

The work spans hardware-adjacent software, model research, simulation platforms, and technical interfaces.

The portfolio is organized around engineering surfaces that reinforce each other rather than around a generic skill list.

Robotics software

Robot code, operator workflow, subsystem architecture, and match-pressure reliability.

AI / ML

Small language models, behavioral modeling, face recognition, and practical training pipelines.

Computer vision

Tracking, OCR-assisted workflows, camera systems, and video-to-field analysis for robotics review.

Simulation

Deterministic runtime design, replayability, autonomous validation, and real-code compatibility.

Tooling / infrastructure

Planning tools, strategy backends, templates, and the shared engineering surfaces around the code.

Interface design

Technical UIs that stay readable under pressure instead of treating design as an afterthought.

Featured work

Selected projects that show range, depth, and technical intent.

These projects were chosen to show robotics leadership, ML research, deterministic simulation, and vision tooling without flattening everything into a generic gallery.

Transformer block diagram from the personalized SLM project.
AI / ResearchResearchIndependent research track

Designing and Training a Personalized Small Language Model Capable of Human Behavioral Replication

Independent-study SLM work spanning baseline transformers, from-scratch modeling, optimizer and tokenizer experiments, and a local personalized agent track.

Personal research and implementation project.

Domains

AI / MLResearchSystems design
Python / PyTorch / Transformer architectures
RoboticsRobotics software

FRC Team 1884 Season 2026 Software Stack

RoboticsTeam repoActive team codebase

FRC Team 1884 Season 2026 Software Stack

Public team robot code with operator board UI, state-first subsystem architecture, characterization workflows, and simulation-ready autonomous tooling.

Team repository. The site represents leadership, codebase responsibility, and contribution context rather than claiming sole authorship.

Domains

Robotics softwareSystems designTeam engineering
Java / WPILib / AdvantageKit
Autonomous ToolingRobotics tooling

PathPlanA

Autonomous ToolingTeam repoActive team toolchain repo

PathPlanA

Flutter planner app for REBUILT autonomous authoring that exports deploy-backed auto libraries into the FRC 1884 robot workflow.

Team repository. Included as part of the FRC1884 software ecosystem, not as upstream authorship of PathPlanner itself.

Domains

Robotics toolingInterfacesSystems integration
Dart / Flutter / Deploy library contracts
SimulationSimulation

GriffinSim

SimulationTeam repoDeterministic simulator monorepo

GriffinSim

Deterministic FRC-compatible simulation monorepo centered on replayability, fixed-step runtime design, and strict contracts.

Team repository represented as part of the broader FRC1884 simulation and tooling surface.

Domains

SimulationSystems designRobotics tooling
Java / Gradle / Deterministic runtime design
Alliance heatmap output from the FRC Heatmap project.
Vision ToolingTeam repoLocal web app / public repo

FRC Heatmap

Local web app that converts wide-shot FRC match footage into field-space robot heatmaps, alliance overlays, and exportable tracking data.

Public FRC1884 repo with local artifacts available during the audit.

Domains

Computer visionRobotics toolingWeb apps
React / TypeScript / FastAPI

Leadership

Technical credibility is stronger when it comes with shared responsibility.

The team context matters because it shows work done inside a real collaborative environment, with deployment pressure, operator needs, and public code surfaces.

Current

GitHub / Software Leadership

ASL Robotics / FRC Team 1884

Operating inside a collaborative robotics software environment where repository structure, review quality, planning tools, and operator-facing systems affect real hardware and real match preparation.

Responsibilities

  • Help shape repository organization and code workflow around the team’s software work.
  • Contribute across robot code, operator-facing tooling, simulation work, and public engineering utilities.
  • Translate technical complexity into interfaces and processes other team members can use under time pressure.
  • Treat GitHub as an engineering surface: code review, structure, versioning, and shared ownership matter as much as the code itself.

Public evidence

  • Public org repositories now span Season2026, PathPlanA, GriffinSim, FRCHeatmap, frc1884-strategy, frc1884-scouting, and attendance tooling.
  • PRE-Season2026 and season2025 public READMEs both explicitly credit PathPlanner + PathPlannerLib, Choreo/ChoreoLib, and maple-sim as part of the 2026-era stack.
  • Season2026 docs explicitly preserve PathPlanA autos through deploy and operator-board selection flows.

GitHub footprint

The public repo surface is broader than the featured case studies.

Beyond the main projects, there is a wider layer of strategy software, scouting systems, operations automation, and earlier season code that helps ground the portfolio in real ongoing work.

How I work

I care about how systems are structured, how they are used, and whether they survive contact with reality.

That usually means thinking past the main algorithm and into the deployment path, review flow, debugging surface, and interface quality.

Systems before features

I care about boundaries, data flow, and operator workflow before I care about adding one more feature checkbox.

Build for real constraints

Robotics, simulation, and ML work only get interesting once timing, reliability, and debugging pressure enter the room.

Design is part of the engineering

Readable interfaces, good defaults, and visual clarity matter when a system has to be used quickly and correctly.

Credibility comes from implementation

I prefer concrete architecture, working artifacts, and clear technical tradeoffs over vague claims about innovation.

Contact / links

GitHub is the clearest place to see the code surface right now.

The site is built to make future additions easy. Public writeups, resume links, demos, or papers can be added without changing the visual structure.