Nadir Trapsida Email & Phone Number

Montreal, Qc, Ca

Montreal, Quebec, Canada

Montreal, Quebec, Canada

Montreal, Quebec, Canada

• Increased customer engagement by implementing one of the first veterinary assistance conversational agents based on a RAG, thereby enhancing customer satisfaction and loyalty.• Designed a production deployment process for the RAG to improve system stability and reduce inference errors.• Created a predictive model using IMU data to detect dog activities with 97.5% accuracy, improving the understanding of canine behavior and needs.• Developed backend code to connect all parts of the IoT project, including the database, mobile application, and AI models, significantly reducing project integration time and increasing overall project efficiency.• Efficiently managed and maintained multiple Microsoft Azure servers, resulting in better scalability and reduced operational costs.• Optimized CI/CD pipelines to reduce deployment time by 3 hours and increase deployment reliability, thereby enhancing the development team's efficiency.• Designed a mobile application prototype, leading to a successful presentation to investors and VCs, securing multiple investments.• Managed a team of developers and interns, supervising up to 6 individuals per time, resulting in improved project outcomes and strengthened management skills.

•Developed a 3D simulation of a bus with visual rendering from multiple angles using the CARLA simulator, enhancingthe testing of autonomous driving algorithms.• Improved the object detection model on the CARLA simulator client to accurately identify vehicles, resulting in a 13%increase in detection accuracy and enhanced safety within the simulation environment.• Restructured and integrated numerous segments of a large autonomous vehicle research project, significantlyimproving project organization and contributing to a more organized and efficient project delivery.• Successfully collaborated with NovaBus to prepare the integration of the solution and algorithms into a real bus.

Canada

• Optimized and maintained CI/CD scripts using Makefiles and GitHub Actions, reducing deployment and testing time,thereby enhancing overall efficiency and productivity.• Increased code coverage by 20% through the development of unit tests, ensuring code robustness and minimizingpotential errors.• Discovered a major flaw in a company product, preventing potential data breaches and protecting user privacy.

Montreal, Quebec, Canada

• Improved speed of rendering of the webapp through optimization, resulting in a more responsive user experience and increased user satisfaction.• Established separation of production and staging environments, reducing the risk of potential errors, leading to smoother deployment processes and increased system stability.• Designed and implemented a robust authentication infrastructure with jwt, incorporated decathlon Auth api for a hackathon, and established a refined permission system for controlled access to decathlon AI team apis

Montreal, Quebec, Canada

• Developed a machine learning model with 95% accuracy to classify text related to the Decathlon brand or sport,enabling more effective brand monitoring and better understanding of social media trends.• Created a sentiment analysis machine learning model using tweets and Facebook texts to analyze user sentimenttowards the Decathlon brand, leading to more informed brand decision-making.• Used geometry and facial detection models to design a smile detector in images, ensuring higher quality imagesselected by coaches on social media, increasing customer attraction.• Maintain and Improve a full-stack web application used by multiple teams for social listening to improve customer understanding

Sherbrooke, Canada

• Implemented Machine Learning models to efficiently detect small targets and objects in drone-captured images, streamlining processes for competitive events.• Adapted a pre-trained model through fine-tuning to tailor it for our specific use case, optimizing its performance and accuracy.• Implemented sophisticated machine learning algorithms within Docker containers, ensuring smooth integration and efficient deployment for streamlined processes in competitive events.• Engaged in collaborative research with a post-doctoral student, combining expertise to advance knowledge and tackle challenging scientific problems.• Presented at multiple events to recruit new students and secure sponsorship, contributing to the growth and sustainability of the club.

Montreal, Canada

• Collaborated on enhancing object detection models by integrating pretrained models from the "Open Images" dataset.• Designed a personalized sports recommendation system leveraging user location, purchase history, and expressed interests, aimed at enhancing user engagement and satisfaction, potentially leading to increased customer loyalty and sales.• Designed and implemented scraping scripts to efficiently gather images from diverse online sources, facilitating the analysis of popular sports on social media platforms and seasonal trends.• Optimized AI server endpoints with protobufs to reduce prediction query time by 83%, from 12 to 2 seconds, positively impacting operational efficiency by speeding up request processing.• Developed a machine learning model to provide crucial statistics for a marathon (colors used, brands...), enabling precise analysis to guide the product team.• Created a social media monitoring tool to analyze the performance of the Decathlon brand on various platforms (Twitter,Instagram, Facebook, and Google Trends), allowing the team to stay up-to-date on brand trends.• Implemented a word vector integration method to create a complete application (frontend, backend, database) facilitating the search for relevant images from social platforms based on text. Achieved 90% image search relevance, used for better customer understanding.

Boisbriand Qc, Cananda

•Designed and implemented server prototypes for clients, increasing customer satisfaction and successful adoption ofnew server technologies.• Conducted benchmarks on Tesla P4, T4, Vega-4000 graphics cards for compatibility with 5 server families, ensuringseamless integration and functionality.• Automated network, performance, and power consumption tests using Python and Bash scripts, reducing testing timeby 4 hours and enabling faster product iterations.• Developed a web application to manage server maintenance, reduce power consumption, and enable remote serverroom management, leading to more eco-friendly and efficient operations.

Boisbriand Qc, Cananda

• Worked in Agile methodology, enhancing project adaptability and team collaboration.• Performed software, thermal, and mechanical tests, ensuring the reliability and efficiency of the system.• Certified servers with different operating systems, including Ubuntu (16 and 18), Red Hat (7.8), and Windows Server (2016, 2012 R2, and 2019), ensuring compatibility and performance.• Configured network switches, optimizing network performance and connectivity.

Sherbrooke, Canada

• Designed and developed a car charger system using a microcontroller, ensuring compliance with automotive safety and efficiency standards.• Researched and tested various microcontrollers to determine the most suitable option based on performance, power consumption, and integration ease.• Selected the optimal microcontroller, leading to improved charging efficiency and system reliability.• Designed and optimized buck, boost, and buck/boost converter circuits, achieving high efficiency and reliable voltage and current regulation.• Prepared detailed presentations for other teams, outlining the design process, microcontroller selection, and circuit implementation, facilitating knowledge sharing and collaboration.

Master'S Degree, Artificial Intelligence

Bachelor'S Degree, Computer Engineering

Preparatory Certificate For Undergraduate Programs In Engineering, Science And Health, Computer Engineering

Baccalauréat, Sciences With Mathematics Specialisation, Admitted With Distinction

Activities and Societies: School soccer team