Zeyd A.

Key Skills

Software

Other

Top Subject Matter

Dietary management and recipe/ingredient extraction using computer vision and VLMs

Programming education and algorithm instruction

Intro programming and robotics for children

Top Data Types

Image

Top Task Types

Fine-tuning

Freelancer Overview

Lead AI Software Engineer | Fridge Inventory AI Startup. Brings 2+ years of professional experience across complex professional workflows, research, and quality-focused execution. Core strengths include Other. Education includes Master of Science, University of Warwick (2026) and Bachelor of Science, Coventry University. AI-training focus includes data types such as Images, Computer Code, and Programming and labeling workflows including Fine-tuning, Computer Programming, and Coding.

Labeling Experience

Online Coding Tutor | Sherpa

Other

Delivered 1-to-1 online tutoring focused on programming, algorithms, and computing fundamentals. Assisted learners with coding concepts that support building and understanding AI-related systems. Adapted teaching to each student’s progress and learning needs. • Mentored students in programming and algorithms • Supported computing fundamentals for practical learning • Guided problem-solving and debugging approaches • Provided individualized tutoring sessions

2025 - Present

Coding Club Instructor | Code Crazy

Other

Instructed primary school students through interactive Python and robotics programming lessons. Focused on engaging, hands-on learning that builds coding literacy relevant to later AI/automation exposure. Designed lesson delivery to match children’s skill levels and learning pace. • Delivered interactive Python programming lessons • Taught robotics programming concepts • Led structured hands-on activities for learners • Adapted instruction for primary school students

2025 - 2026

Lead AI Software Engineer | Fridge Inventory AI Startup

OtherFine-tuning

Fine-tuned a custom vision-language model to transform raw user images into strict JSON ingredient lists, reducing LLM hallucinations in structured outputs. Engineered the model workflow to perform image-to-text extraction with enforced JSON formatting for downstream use. Worked on end-to-end pipelines where vision inference was isolated from lighter CPU database tasks. • Fine-tuning for vision-language extraction • Generated strict JSON outputs from images • Reduced hallucinations via constrained output format • Integrated with asynchronous streaming cloud API workflows

2026

Education

Work History