
Yassir CHERDOUH
AI/Software Developer
About Me
I'm a 3rd-year Artificial Intelligence and Data Science student at ENSIA. I build web applications and train systems to think. I'm passionate about solving real-world problems through data-driven solutions and intuitive digital experiences. I'm always looking to collaborate, learn and grow.
Professional Focus
Applying Machine Learning and Data Science to solve complex problems and extract meaningful insights
Building web and software solutions that are intuitive, scalable, and impactful for real-world use
Leading and contributing to innovative projects and researches across AI, Open-Source, and tech media
Languages
Qualifications & Expertise
Professional Certifications

Fundamentals of Deep Learning
NVIDIA · 2025
ID: 1tO0Ys3ITkGJkXM3sgBKrQ
Education

Master of Science: Artificial Intelligence & Data Science Engineering Degree
The National Artificial Intelligence School (ENSIA) · 2023-2028 (Expected)
BAC Diploma, Speciality: Maths, Average: 18,65/20 (Rank: 1, Bejaia)
Hafsa Oum El Mouminine High School · 2023
Areas of Expertise
Experience
Full-Stack Engineer & System Architect
I am part of the team building a centralized, ministry-scale platform that lets the Ministry of Higher Education and Scientific Research (MESRS) securely distribute official documents to its subordinate institutions, enforce fine-grained access control, and search the institutional archive through AI-powered semantic search.
- Authored the Software Requirements Specification (SRS) and co-designed the 4-layer system architecture (Web, Application, AI/ML, Data).
- Integrated single sign-on (SSO) with the ministry identity provider, including automatic user provisioning and role-based access gating.
- Managed the full document lifecycle: versioning, automatic archival, auto-generated reference numbers, tagging, and a Draft → Active → Archived state machine.
Internships
Contributing to a national-scale AI assistant that guides Algerian baccalaureate students through the annual university enrollment process. The assistant launches publicly right after the baccalaureate results to help students navigate official admission information in natural language.
- Led the team and owned the engineering setup end-to-end: CI/CD pipelines, automated deployment, and Dockerized environments for consistent local-to-production parity.
- Designed and shipped a multi-agent RAG system grounded strictly in the ministry's official knowledge base (Arabic / French / English), using multilingual embeddings so a question in one language matches an answer in another.
- Built a hybrid retrieval strategy: hand-curated bilingual article summaries answer first, with OCR'd passages from the official circulaire as a fallback tier, plus a measured confidence threshold that makes the assistant refuse rather than hallucinate when nothing relevant is found.
- Implemented OCR-based document intelligence: the official circulaire's Arabic text wasn't machine-readable, so I built an OCR ingestion pipeline that converts it into clean, retrievable text with a vision model. I also built a BAC-transcript attachment pipeline (Ghostscript PDF→PNG + vision-model reading) that lets students upload official documents directly in chat.
Technical Skills
Languages
Frameworks & Libraries
AI/ML
DevOps & Databases
Leadership & Community
HR Head
I lead the HR department of the largest scientific club at ENSIA and the Sidi Abdallah Technological Pole, a community of more than 300 members, and take ownership of several of its key projects.
- Coordinated and supported the delivery of 6+ technical events, including dedicated event websites and a custom QR-based check-in system that streamlined attendance tracking.
- Led the FOSS Flash open-source newsletter and represented the club in external media, including ETCODE 3.
- Developed the club events' websites (Bitcamp4, ETCODE4, EXAI..) and internal automation tools (Attendace Check-in System, ETC Email Sender..)
Hackathons & Competitions
FORSAtic Hackathon
3rd Place / 64 TeamsAlgérie Telecom
Built an AI system that classifies customer complaints from social media and call centers automatically, giving support teams a fast way to surface and prioritize urgent cases.
Challenges
Challenge 1: Social Media Comments Classification
Classify customer comments from social media into 9 distinct categories
Challenge 2: Call Center Tickets Classification
Classify customer support tickets from call centers into 6 distinct categories
Solution Highlights
- Classified customer messages by their nature (complaint, outage report, request for assistance, acknowledgement) and by urgency level
- Gave support teams a clear way to surface and prioritize critical complaints
- Tracked satisfaction trends and recurring topics through visual dashboards updated in real time

Certificate
This was supposed to be the picture of the winning ceremony, but no closing ceremony was held.I officially lost hope...
Winning Moment
Let's build something
I'm open to internships, collaborations, and interesting problems. Reach out and I'll get back to you.
Location
Bejaia / Algiers, Algeria
Response Time
Usually within 24 hours




