AI Developer PortfolioExplore

Yousif SalmanAI Developer

Available Now

Welcome to my creative playground! I'm Yousif Salman

I craft intelligent AI solutions for businesses, and develop innovative applications that leverage the power of artificial intelligence.

About Me

AI Developer passionate about building intelligent solutions that solve real-world problems

Who I Am

I'm an AI Developer with a passion for creating intelligent systems that make a real impact. Currently pursuing my Bachelor of Science in Computer Science at Wilfrid Laurier University, I've worked on cutting-edge AI projects that have improved efficiency, accuracy, and user experience across various industries.

From building transformer-based LLM systems to developing machine learning models for clinical decision-making, I thrive on solving complex problems with innovative AI solutions.

What I Do

AI Development

Building intelligent systems with LLMs, transformers, and deep learning models

Full-Stack Development

Creating end-to-end solutions with modern frameworks and APIs

Research & Innovation

Contributing to academic research and presenting findings at conferences

Skills

Languages

PythonJavaScriptC++JavaGoSQLBashHTML/CSS

Frameworks

React.jsNext.jsNode.jsFastAPIPyTorchTensorFlow

Libraries & Tools

NumPyPandasscikit-learnHugging Face TransformersOpenCVGitLinux

Published Research

Contributing to AI research in healthcare and clinical applications

Improving Patient-Clinical Trial Matching Using Convolution Neural Networks

Yousif Salman, Emad Mohammed, and Cassandra HuiThe 38th Canadian Conference on Artificial Intelligence (CAIAC 2025)2025

Developed a neuro-symbolic AI approach combining fine-tuned LLMs with symbolic reasoning to improve patient-clinical trial matching. The method enhances interpretability and accuracy by validating LLM and cosine similarity outputs through additional layers of symbolic reasoning, providing more reliable and transparent trial-matching results.

Read Paper

Evaluating Missing Data Imputation Methods for Esophageal Cancer Quality of Life Research

Yousif Salman et al.medRxiv2025

Comprehensive evaluation of seven imputation methods for handling missing data in esophageal cancer patient-reported outcomes. Compared methods including MICE, VAE, DAE, BPCA, deep autoencoders, SoftImpute, and KNN across multiple validation metrics to provide evidence-based recommendations for quality-of-life research.

Read Paper

Work Experience

Building AI solutions that make a real impact

AI Software Developer

Laulima Consulting Inc.
Remote
Sep 2025 – Present
  • Engineered an AI system using transformer-based LLMs with embedding-based similarity scoring, automating job description creation and evaluation for over 3,000 roles.
  • Built custom conversational AI tools with FastAPI and LangChain, enabling natural, iterative refinement of job descriptions and streamlining HR collaboration.
  • Developed automated scoring and SHAP-inspired explainability modules with rule-based logic, improving fairness and transparency while increasing reviewer satisfaction by 25% and achieving 90% reviewer agreement.
  • Optimized model performance and integrated the system into Laulima's HR workflow, cutting review time by 40% and boosting classification accuracy by 15%.
PythonFastAPILangChainLLMsTransformers

Coding & Math Instructor

ICan Education
Waterloo, ON
Nov 2023 – Nov 2025
  • Integrated three new AI and real-world application modules from scratch, seamlessly integrating them into the existing computer science curriculum.
  • Tracked student performance across coding and interdisciplinary projects, providing targeted feedback that led to an average score improvement of 20%.
PythonJavaScriptTeachingCurriculum Development

AI Developer & Research Assistant

McGill University
Remote
May 2025 – Aug 2025
  • Developed machine learning models for missing data imputation, increasing data completeness by 45% and improving stability across timepoints, by combining statistical baselines with deep learning methods.
  • Enhanced clinical decision-making accuracy, reducing prediction error on key outcomes by 18%, by building an end-to-end system that integrated imputation pipelines with predictive analytics.
  • Improved real-world usability for clinical teams, shortening analysis workflows by 30%, by creating a user-friendly interface and validating outputs through multi-site clinician feedback.
PythonPyTorchDeep LearningClinical AI

AI Developer

Wilfrid Laurier University
Waterloo, ON
Sep 2024 – Dec 2024
  • Led the development of an AI model that matched patients to clinical studies, achieving a 92% reduction in recruitment time through automated eligibility screening.
  • Built the full end-to-end pipeline including data preprocessing, model architecture, and deployment, improving matching accuracy by 6% compared to traditional recruitment methods.
  • Presented a research paper at a top Canadian AI conference, sharing insights on intelligent patient–study matching with over 100 attendees.
PythonMachine LearningResearchData Science

Education

Building a strong foundation in computer science

Bachelor of Science, Computer Science (Co-op)

Wilfrid Laurier University
Waterloo, ON
Expected Aug 2026

Pursuing a comprehensive computer science degree with a focus on artificial intelligence, machine learning, and software engineering. The co-op program has provided valuable real-world experience in AI development and research.

Get In Touch

I'm always open to discussing new projects, creative ideas, or opportunities to be part of your vision.

Let's work together to bring your ideas to life