Eineje A.

Key Skills

Software

Don't disclose

Top Subject Matter

LLM evaluation for multi-robot collaboration and robot control

3D LiDAR mapping and point cloud visualization

Control systems for aircraft pitch stabilization

Top Data Types

3D Sensor

Image

Top Task Types

Mapping

Classification

Freelancer Overview

Research Assistant, Michigan Technological University (LLM/VLM testing and evaluation). Brings 5+ years of professional experience across complex professional workflows, research, and quality-focused execution. Core strengths include Don't disclose, MOLA, and LAZ. Education includes Doctor of Philosophy, Michigan Technological University (2026) and Master of Science, Michigan Technological University (2026). AI-training focus includes data types such as Computer Code, Programming, and 3D Sensor and labeling workflows including Evaluation, Rating, and Mapping.

Labeling Experience

Advanced Programmable Logic Controller, MTU (AI vision + PLC conveyor defect routing)

Built an AI vision plus PLC conveyor system intended to classify defects and route parts for accept, rework, wrong part, or scrap outcomes. Programmed PLC logic for reliable high-speed sorting with fault handling such as camera timeouts and jam detection. Implemented SCADA monitoring and KPI dashboards to track cycle time, downtime, and defect counts for operational feedback. • Designed defect classification workflow integrating AI vision with PLC routing.• Implemented PLC state machine logic including fault handling and reject verification.• Developed Ignition SCADA monitoring with defect counts and cycle/downtime metrics.• Produced structured tag-based dashboards for performance tracking.

2026 - Present

3D LiDAR Campus Mapping (Project/Demo)

3D SensorMapping

Built an end-to-end 3D LiDAR mapping pipeline for an interactive browser-based point cloud map. Transformed raw LiDAR bag data into a motion-corrected, metric representation using LiDAR odometry and scan-to-scan registration. Processed point clouds with filtering and compression to support accurate visualization and reduced noise. • Generated a motion-deskewed metric map from LiDAR data.• Applied voxel occupancy filtering to reduce dynamic-object noise.• Optimized point cloud parameters (crop bounds, voxel size, elevation coloring, density) for web deployment constraints.• Compressed and tiled point clouds using LAZ and PotreeConverter for efficient interactive streaming.

2026 - 2026

Research Assistant, Michigan Technological University (LLM/VLM testing and evaluation)

Don't disclose

Conducted evaluation of LLM behavior for multi-robot collaboration by testing Brick-GPT and analyzing its prompt limitations. Assessed rejection patterns and stability constraints to understand model performance under different prompting conditions. Documented findings to support further improvements to LLM-based control approaches.• Tested Brick-GPT and analyzed prompt limitations.• Studied rejection patterns and stability constraints.• Benchmarked solver-related runtime/output differences to inform model reliability.• Produced implementation documentation to enable reproducible AI evaluations.

2025 - 2026

Control System- Aircraft Pitch Control (Using PID- MATLAB)

Classification

Modeled and validated an aircraft pitch control system and compared open-loop versus closed-loop behavior. Tuned and analyzed PID controller dynamics using pole-zero analysis and step response evaluation. Verified stability improvements by observing settling to target values instead of drift. • Developed a MATLAB pitch control model from given transfer functions.• Designed a PID controller and applied unity feedback for stable tracking.• Performed pole-zero analysis to study controller influence on system dynamics.• Compared step responses to evaluate stability and tracking performance.

2025 - 2025

Education

Work History