CS @ Northeastern University • AI & Software Engineering • Data-driven Problem-solving • VFX/Computer Vision • Football Analytics
📧 Email · 🔗 LinkedIn · 💻 GitHub
Co-op Availability: January – June 2026
| AI + Data LLM → SQL, analytics pipelines, Earth observation |
Computer Vision Segment Anything + video mask tracking |
Full-Stack React/Expo, Streamlit, FastAPI, scalable apps |
- AI & Data Systems: Build end-to-end analytics apps that pair LLMs with robust SQL views/pipelines for fast insight generation.
- Vision for VFX: Integrate Segment Anything into interactive tooling to create, refine, and track masks across frames.
- Product & Teams: Lead small teams, run sprints/code reviews, and ship usable features with clear impact.
Frameworks & Tools: React/Expo, Streamlit, FastAPI, Unity, Unreal Engine, Git, Linux, AWS
ML/DS: TensorFlow, scikit-learn, NumPy, Pandas, Matplotlib, Tableau, Mistral
- Northeastern University, Khoury College of CS — B.S. Computer Science
- Interests: AI in sports & business, robotics, game dev, body-building, FC Barcelona, FIFA/Halo, Formula 1
Interactive video object tracking & segmentation app built on Segment Anything. Enables point-and-click mask creation and tracking across frames for VFX workflows.
- Streamlit UI for fast selection, review, correction
- Multi-object support and mask export
Cross-platform React/Expo app with a lightweight recommendation engine, designed for usability at scale.
- Mobile-first UX, clean component architecture
- Team sprints, code reviews, and CI basics
AI system over multispectral Earth observation data. LLM-to-SQL querying + curated SQL views for pre/during/post-event analysis and rapid insights for non-technical users.
- Natural-language queries → structured analytics
- Causal and temporal breakdowns across phases
ML models (KNN → DQN) on player datasets to explore scouting, health, and tactical decision-making; highlights trade-offs between interpretability and predictive power.
- End-to-end Python pipeline and benchmarks
- Clear metrics & visualizations for insight
- 🎬 Potentially reduced prep time by 20% in VFX pipelines at DNEG by integrating Segment Anything into Streamlit tooling.
- 🌊 Improved flood prediction accuracy by 25% with an AI-powered platform at Atos (Eviden), enabling 100+ daily queries from non-technical users.
- 📱 Boosted engagement by 30% leading a 5-member team to deliver a food delivery app with recommendation engine at OASIS Dev Club.
- ⚽ Applied ML to predict outcomes for 500+ football players, benchmarking models from KNN to Deep Q-Networks.