Looking for a job

bedis ammar

@bedisammar

Deep learning researcher at ESPRIT, reproducing Show, Attend and Tell and benchmarking four image-captioning architectures.

Tunisia

Message

What I'm looking for

I want to move from applied ML into research on scalable, modular and verifiable deep learning. I'm seeking a six-month internship starting January 2027, remote from Tunis.

At ESPRIT, my research project reproduces Show, Attend and Tell in PyTorch and benchmarks it against three other image-captioning approaches under a shared evaluation protocol. The trained Transformer decoder outperformed the reproduced baseline on METEOR and CIDEr, and I examined model behavior with attention visualizations, Grad-CAM, error analysis, and robustness testing.

At Sagemcom Software & Technologies, I designed an eight-agent LLM/RAG coverage engine to audit Jira test cases and suggest requirement links; technical validators approved it. I also automated Jira/Xray and RETRACK reporting, reducing the process from about two hours a day to under two minutes. I want to move from applied ML into research on scalable, modular, and verifiable deep learning, and am seeking a six-month internship starting January 2027.

Experience

Work history, roles, and key accomplishments

AI & Software Engineering Intern

Sagemcom Software & Technologies

Jul 2026 - Aug 2026 (1 month)

Designed an 8-agent LLM/RAG coverage engine that audits 1,350 Jira test cases and proposes requirement links ranked by confidence. Automated JIRA/Xray and RETRACK reporting with Flask and Angular, reducing reporting time from 2 hours to under 2 minutes.

LLM RAG Multi Agent Systems Jira Xray Flask Angular Automation

AI Engineer

Laboratoires VITAL

Jan 2025 - Jan 2026 (1 year)

Built a product-recommendation agent over a 511-product catalogue using FAISS and LLaMA 3 70B, and developed CRM report structuring and social-media content pipeline modules. Engineered product features from 22 to 174 and created a sentiment lexicon for French, Tunisian, and Arabic reviews.

Faiss LLaMA 3 RAG Recommendation Systems CRM Sentiment Analysis Data Engineering

Networks & Telecom Intern

Tunisie Telecom

Jul 2025 - Aug 2025 (1 month)

Analyzed fibre-network deployment data during a summer internship at Tunisie Telecom.

Education

Degrees, certifications, and relevant coursework

ESPRIT — School of Engineering and Technology

Engineering Degree (Master's-level), Computer Science, Data Science & Artificial Intelligence

2024 -

Grade: GPA 13.85/20 (French scale)

Pursuing a Master's-level Engineering Degree in Computer Science with a focus on Data Science and Artificial Intelligence. Final-year research project (PFE) to be completed by November 2027.