Disclaimer: This is not an AI template nor an AI-generated website! Thanks for visiting, this portfolio is a work in progress, In the meantime, feel free to explore and visit my LinkedIn profile to see more of my professional background and experience. Expected completion date: December 2026.
Disclaimer: This is not an AI template nor an AI-generated website! Thanks for visiting, this portfolio is a work in progress, In the meantime, feel free to explore and visit my LinkedIn profile to see more of my professional background and experience. Expected completion date: December 2026.
Ndeye Awa Salane
MSc Computer Science Candidate | Machine Learning & R&D
I'm a final-year Master’s student in Computer Science at the University of Geneva, after my engineering degree in Computer Science and Telecommunications at École Polytechnique de Thiès (EPT) , which I joined in 2020 through a national selection process.
I'm expected to graduate in September 2026 and am seeking full-time Machine Learning Engineering roles starting Fall 2026, or related positions in the field. My interests include Machine Learning, Deep Learning, Reinforcement Learning, Information Retrieval, and Generative AI. I am also curious about Metaheuristics, Automata Theory, and enjoy building well-designed UX/UI experiences.
Beyond tech, I love cooking, crochet, sketching, working out (HIIT, Core, Strength) and I'm a karate green belt. And yes, stay tuned! Recipes might show up here too.
Technical Skillset
professional Experiences
.png "Roche"){kind=logo}
July 2025 - April 2026
- Engineered a high-throughput patent data extraction pipeline, with batching, chunking, and parallel execution, reducing manual review from 5 hours to under 2 minutes per patent.
- Benchmarked LSI/BIR models; built neural retrieval system with semantic search to narrow 300+ patents into top-k shortlists for review.
- Collaborated with domain experts and legal teams to define and refine features for patent relevance and compliance.
- Worked on column classification with GEPA prompt optimization to provide users with mapping of oligo sequences to patent readouts.
November 2024 - April 2025
- Fine-tuned NLLB-200 using LoRA for low-resource Wolof-French translation on a 10,000+ parallel sentence dataset.
- Achieved +70% BLEU and +34% ROUGE improvement over prior baselines for the French to Wolof direction.
- Managed end-to-end training loops and metric tracking then deployed the final model on the Hugging Face Hub.
January 2024 – June 2024
- Architected a multi‑agent system routing queries across 5+ benchmarked LLMs, optimizing for Text-to-SQL, data visualization generation, non-database requests and overall inference latency.
- Achieved 96% Text-to-SQL accuracy on proprietary schemas using CoT prompting, few-shot learning, and iterative evaluation and validation.
- Built a data assistant that delivers records and visual insights and enabled 25% faster data access for all users regardless of their technical proficiency.
September 2023 – January 2024
- Built a scalable data validation framework enforcing 30+ custom rules on multi-GB daily batch datasets in the data lake before warehouse ingestion.
- Implemented validation jobs (cron-based) triggered on daily data ingestion, generating reports with data quality metrics such as schema violations, completeness scores, and column-level anomalies.
- Improved data quality by 40% by intercepting and filtering invalid records for more reliable downstream analytics and other data-driven workflows.
- Developed and deployed a self-service UI for teams to create and run dataset-specific quality checks.
October 2022 - December 2022
- Designed and implemented functional and structural software models using UML diagrams.
- Developed backend and frontend applications, implemented containerization for efficient deployment.
Certifications
Projects I worked on
Some Topics I wrote about
[Coming Soon...]