Hey, I'm Jayansh Bagga.
I’m a Computer Science student at Western and Founder & Technical Lead of Savify, a privately-backed fintech startup. I build scalable systems across software, AI, and finance from LLM infrastructure to backend applications. I also believe that computer science students can take a cue from the finance world, sharp code deserves a sharp suit.
About Me
I’m a Computer Science student at Western University, building at the intersection of software, AI, and finance. I’ve been working on real-world systems—from LLM-based infrastructure to scalable backend applications. I’m especially interested in building products that are not just technically strong, but genuinely useful.
When I’m not coding, I’m usually expanding my vocabulary, picking up new languages, sampling colognes and cognac, or just touching grass to balance it all out.
Currently Working On
- Working as a Data Analyst at RBC in Toronto, supporting Personal Banking analytics and reporting within the Decision Management team.
- Wrapping up Savify after running out of runway, and exploring what's next — including deeper work on Project Kalson, designing systems that learn and solving problems that actually hurt.
- Developing my personal brand and investment portfolio while pursuing bigger professional opportunities.
Projects
Savify is a privately-backed fintech platform focused on goal-based savings, combining machine learning insights, financial analytics, and gamified user engagement. Currently at 150+ registered users with a live platform and mobile app in development, scaling toward 500+ users post-launch.
Built a FastAPI semantic cache for Boardy ($8M AI startup) using vector embeddings and cosine similarity to reduce LLM calls by 60% and latency by 40%. Designed as a modular caching engine for AI applications.
Reflecta is an AI-powered reflection system built during UofTHacks 13 that helps users analyze personal thoughts and writing using large language models. The project explores how AI can transform raw journaling into structured insights and pattern detection for self-awareness.
Education
B.Sc – Major in Computer Science
Skills
Work Timeline
Data Analyst
Royal Bank of Canada
Sept 2026 - Dec 2026
Software Engineering Intern
Flynn Group of Companies
May 2026 - Sept 2026
AI Engineer
Outlier AI
Sept 2024 - May 2025
Experiences
Data Analyst
Royal Bank of Canada • Sept 2026 - Dec 2026
- Building automated SQL and Python reporting pipelines within the Decision Management team to support data-informed decision-making for RBC's Personal Banking division (16M+ customers).
- Partnering with cross-functional teams across Risk, Marketing, and Product to translate business questions into data models, optimizing ETL/SQL workflows and reducing data processing time by ~25%.
- Contributing to reporting and analytics initiatives influencing $500M+ in managed portfolio value, cutting manual reporting turnaround by ~35% across the Personal Banking analytics team.
Software Engineering
Flynn Group of Companies • May 2026 - Sept 2026
- Tasked with modernizing a customer-facing platform tracking $100M+ in construction projects, built and shipped 25+ Angular/TypeScript features end-to-end from UI to backend integration, improving workflow efficiency and user experience for daily project-management users.
- To eliminate slow, error-prone manual approvals across enterprise systems, designed and deployed REST API endpoints in .NET Core and SQL that automated multi-stage approval workflows, cutting processing time 30% and removing manual data-entry errors.
- Embedded in a cross-functional Agile team of PMs, engineers, and business stakeholders, drove features through the full SDLC across the project lifecycle, contributing to a platform managing 500+ commercial construction projects across North America.
AI Engineer (LLM Systems)
Outlier AI • Sept 2024 - May 2025
- Engineered and optimized LLM-based systems using OpenAI APIs and Python, improving output accuracy by ~10–15% across production use cases.
- Built evaluation pipelines with Python and Pandas, benchmarking performance across 1K–5K+ model outputs.
- Scaled NLP workflows on 10K+ text records, supporting enterprise AI deployments and supervised learning tasks.