Tom Bijaoui

Key achievements

• AI Shopping Assistant for Shopify: Designed and built an agentic RAG platform powering AI assistants for Shopify stores, capable of answering customer questions, promoting products and brand identity, and improving first-time visitor conversion rates through advanced retrieval and reasoning strategies.
• E-commerce AI Copilot: Developed an AI agent that continuously monitors merchants' business metrics, detects anomalies in real time and proactively notifies teams through Slack with AI-generated root-cause analyses and recommended actions, significantly reducing reaction time to critical business events.
• Predictive Analytics Platform: Built scalable machine learning pipelines to forecast key e-commerce metrics such as revenue, customer lifetime value (LTV) and churn, enabling merchants to anticipate business trends and make data-driven decisions.
• Customer Feedback Intelligence: Developed NLP models to automatically classify customer reviews, identify recurring product issues and alert merchants when abnormal trends emerge, helping prioritize quality improvements.

Technologies

Key achievements

• Cybersecurity Analytics Platform: Contributed to the development of an internal analytics platform enabling security teams to upload application logs, compute cybersecurity metrics and visualize system behavior through an interactive interface. Implemented data processing pipelines and analytics modules to transform raw logs into actionable security insights.
• Anomaly Detection: Developed an AI-powered anomaly detection system based on unsupervised LSTM models to automatically identify suspicious user activities from application logs. By learning normal behavioral patterns and detecting abnormal event sequences through reconstruction errors, the system generated daily security reports highlighting unusual logins and potential cyber threats, reducing the need for manual log inspection.

Technologies

Key achievements

• Automated Roof Analysis Platform: Designed and developed a production-ready web application that enables users to upload house images and automatically receive optimized skylight placement recommendations. Built the complete solution, including the web application, AI backend and deployment.
• Vision AI & Geometric Optimization: Leveraged Vision Transformer-based models (GroundingDINO and SAM2) to detect roof geometry and obstacles, then developed optimization algorithms to compute available installation areas and recommend the best skylight locations based on spatial constraints.