Francis

Key Skills

Software

Other

Labelbox

Encord

Scale AI

Top Subject Matter

Robotics process control and automated QA evaluation for lab automation

Dynamic system identification and control using sensor time-series

Computer vision multi-object tracking for real-time systems

Top Data Types

Text

Image

Document

Top Task Types

Function Calling

Fine-tuning

Tracking

Prompt + Response Writing (SFT)

Data Collection

Text Summarization

Text Generation

Object Detection

Computer Programming/Coding

Freelancer Overview

Junior Robotics Engineer — production automation and AI-informed evaluation (TRICCA Technologies). Brings 3+ years of professional experience across complex professional workflows, research, and quality-focused execution. Core strengths include Don't disclose, TensorFlow, and OpenCV. Education includes Bachelor of Science, University of Alberta (2026). AI-training focus includes data types such as Computer Code, Programming, and 3D Sensor and labeling workflows including Function Calling, Fine-tuning, and Tracking.

Labeling Experience

AI Fitness Coaching App — LLM integration and adaptive generation (Project)

OtherTextPrompt + Response Writing (SFT)

Designed and integrated an AI coaching application that generates dynamic workout-plan outputs using a large language model API. Implemented a secure backend proxy to route OpenAI calls without exposing client API keys, enabling adaptive plan generation from user performance data. Built a closed-loop data pipeline that persists session metrics and queries prior performance to improve personalization over time. • Architected secure Next.js backend proxy for LLM calls • Integrated OpenAI API for dynamic workout plan generation • Persisted session metrics to PostgreSQL for iterative personalization • Engineered client-side experience with responsive cross-platform UI

2026 - 2026

Junior Robotics Engineer — production automation and AI-informed evaluation (TRICCA Technologies)

Don't discloseFunction Calling

Engineered automation and evaluation logic for an AI-enabled robotics workflow, focusing on generating and validating structured outputs from test pipelines. Implemented acceptance-criteria-based checks that validate positional accuracy and pipetting correctness using automated Python QA flows. Built systems that adapt to logged performance metrics via closed-loop data pipelines, enabling model-informed behavior changes during generation cycles. • Automated Python test pipelines to validate robotic actuator behavior • Used defined acceptance criteria to evaluate correctness • Integrated real-time control modules with a firmware stack • Designed closed-loop metric persistence to drive adaptive outputs

2024 - 2025

Research project — Real-Time Multi-Object Tracking System (Python/TensorFlow/OpenCV)

Tracking

Developed a real-time multi-object tracking pipeline combining detection and motion-state estimation. Implemented a Kalman-filter-based tracker in Python/NumPy to maintain stable identities across multiple simultaneous objects. Integrated a YOLOv5 inference pipeline with a TensorFlow/ONNX backend to achieve strong detection performance on benchmark test sets. • Built Kalman filter tracker for identity maintenance • Fused detections with motion state estimates • Integrated YOLOv5 inference with TensorFlow/ONNX for end-to-end tracking • Optimized latency via batching and adaptive frame skipping

2023 - 2023

Process Control Lab Assistant — ML-based system modeling and control (University of Alberta Research Lab)

3D SensorFine-tuning

Applied machine learning methods to experimental sensor time-series for system identification and control. Reconstructed governing differential equations from sensor-derived signals using a SINDy pipeline implemented in Python. Improved closed-loop behavior and reduced simulation-to-physical discrepancy by applying Kalman filtering, PID tuning, and MPC across benchmark datasets. • Implemented SINDy pipeline for ODE reconstruction • Applied Kalman filtering, PID tuning, and MPC in Python • Prototyped RL and transformer-based models in TensorFlow for dynamic identification • Validated updated parameters using benchmark datasets

2023 - 2023

Education

Work History