Research Engineer - AI, Robotics, & Optimization

Henry Valladares  • Peerlist

Developing a physics-based space game centered around gravitational mechanics and event horizon dynamics. Implemented player movement, collision detection, and interactive gameplay systems. Designed modular scene architecture using Godot’s node-based system. Integrated mechanics such as fuel systems, obstacles, and environmental interactions.

2D Physics-Based Game Development (Godot)

Built an intelligent measurement system for analyzing high-resolution scientific images using computer vision and segmentation techniques. Developed automated feature detection and segmentation pipelines to identify regions of interest and extract meaningful structural data. Designed and implemented custom measurement algorithms to compute real-world dimensions from segmented regions, including calibration from pixel space to physical units. Engineered an interactive visualization interface enabling real-time user interaction, region selection, and measurement validation. Developed a flexible system architecture supporting extensible measurement logic and scalable data processing workflows. Focused on bridging raw image data with real-world physical interpretation, emphasizing accuracy, calibration, and system reliability.

AI-Assisted Measurement & Image Analysis Syst

Designed tricopter and hexacopter drone frames using Fusion 360, focusing on structural layout and component integration. Planned actuator placement, wiring pathways, and onboard electronics configuration for efficient system design. Explored control concepts including thrust vectoring, motor coordination, and stability considerations. Developed a low-cost embedded system architecture for potential real-world implementation.

Multi-Rotor Drone Design and System Architect

I designed and am actively developing OMNI, a local-first AI-assisted engineering platform that transforms natural-language mission requirements into structured, reviewable engineering outputs. OMNI coordinates specialized software agents across robotics, mechanical design, electronics, physics, validation, and technical documentation. Its mission pipeline can interpret an engineering objective, route work to the appropriate specialists, evaluate candidate concepts, calculate reliability metrics, identify risks, and generate traceable artifacts for further development. The platform integrates a FastAPI backend, React/Vite command interface, structured data models, mission knowledge graphs, reliability gates, provenance tracking, and automated testing. OMNI can currently produce starter deliverables for ROS 2 packages, Fusion 360 scripts, KiCad projects, wiring documentation, power budgets, design dossiers, validation reports, and mission summaries. A major focus of the project is ensuring that AI-generated engineering work remains understandable and auditable. Instead of presenting an untraceable answer, OMNI records which components contributed to a result, why decisions were made, what evidence is available, and what the operator should do next. The long-term vision is to develop OMNI into a multidisciplinary engineering laboratory assistant capable of supporting the full workflow from early requirements and system architecture through simulation, electronics, CAD, robotics integration, testing, and fabrication planning.

OMNI - Orchestrated Machine Networked Intelli

Implemented core operating system functionality in C, including thread scheduling and alarm clock mechanisms. Developed system call interfaces and user program support, enabling interaction between user and kernel space. Debugged low-level memory management and process control issues in a Unix-like environment. Validated system behavior through automated testing and kernel-level debugging.

Operating System Development (Pintos)

Implemented distributed communication patterns including ring, broadcast, and scatter/gather using MPI. Developed Cartesian topology communication models for structured parallel processing. Explored GPU acceleration concepts using CUDA and analyzed performance across multiple processes. Measured execution time and evaluated scalability of parallel systems.

Parallel Computing with MPI and CUDA

Developed and launched a 6-DOF robotic arm simulation using ROS2, Gazebo, and RViz to explore robotic control and system behavior. Implemented joint control through ROS2 topic publishing and controller interfaces, enabling real-time manipulation of the robot. Visualized kinematics and system states in RViz, supporting analysis of motion and trajectory behavior. Debugged URDF configurations and resolved controller integration issues within a simulated robotics pipeline.