Uncertainty-aware F1 pit strategy
A strategy agent that decides when to pit under racing uncertainty, with safety-car prediction and measurable validation.
Khush Golani
Open to opportunities
AI tools. Thoughtful systems. Interfaces people enjoy.
Previously at Fractal. Computer Science at Carleton.
Built from curiosity.
Refined through systems.
I’m Khush, a computer science student and software builder who turns messy, high-ambition problems into clear, working products.
My work sits between applied AI, product engineering and interaction design, from uncertainty-aware Formula 1 strategy models to enterprise research agents and the portfolio you are exploring now.
Away from the editor, I run technology operations at a McDonald’s in Ottawa and serve as CFO of Carleton’s CUSEC society.
Studying
Computer
Science
Carleton University, Honours, 2022 to present
Beyond the screen
1st place
Carleton Coding Challenge, 2024. 4th at the Ciena Coding Challenge the same year.
Messy problems into clear,
working products.
Selected work
The thinking, experiments and engineering behind the finished result.
Enterprise context engine
Research and retrieval workflows that make complex enterprise information easier to navigate.
This portfolio
An interactive study in storytelling, graphics performance and where design meets engineering.
Where curiosity
met responsibility.
Engineering
Project Intern
Fractal Analytics (Canada)
Enterprise Context Engine MVP.
- Owned the end-to-end architecture, and built contextualization agents across several AI agent platforms to shape one generalized, cross-platform solution.
- Wrote custom AI skills and agent instructions that extract client needs, match Cogentiq use cases and score relevance.
Tech Lead
Prototyp3
Website and applicant portal for the first developer cohort.
- Built and deployed the website and application portal in 3 weeks with React, TypeScript and Supabase.
- Led full-stack delivery, reviewed code and mentored beginner developers.
Software Team Lead
PixlNext
STO Education Platform.
- Led 4 developers building OAuth sign-in, payments, live classes and a cross-platform desktop application.
- Ran sprints and code reviews, and redesigned the UI and UX with Vite, TypeScript and Supabase.
Leadership
Vice-President
Carleton CUSEC Society
Student software engineering society.
- Elected Vice-President for 2026 to 2027, continuing on from the CFO role.
Chief Financial Officer
Carleton CUSEC Society
Student society finances.
- Prepare budgets and funding applications, and coordinate bus quotes and financial administration for the club.
Operations Technology Manager
McDonald’s Canada
Store technology operations. Promoted to manager in Dec 2025.
- Lead the store’s technology operations and preventive maintenance, and support crew with operational systems and troubleshooting during shifts.
- Recognized as Employee of the Month three times and Shift Leader of the Month.
Tech Lead
Google Developer Student Clubs (GDSC)
Student developer community.
- Hosted multiple networking events and hackathons throughout the academic year for Carleton University students.
Education
Bachelor of Computer Science, Honours
Carleton University
Ottawa, Ontario.
- 1st place at the Carleton Coding Challenge and 4th at the Ciena Coding Challenge (2024), building a dynamic DHCP server under time constraints.
- Won the Best Design Award at Tech-Nol-Hack 2025 for Parity, an anonymous hiring platform that judges talent on merit.
A useful mix
of disciplines.
Evidence
Pit strategy, Context engineApplied AI and machine learning
Models, evaluation and intelligent workflows with clear measures of what works.
Evidence
This portfolioFrontend and creative development
Visual precision, purposeful motion, accessible performance.
Evidence
Context engineBackend and data
Reliable services and structured data.
Evidence
Prototyp3Product and prototyping
Make the problem concrete, build a small version first.
Evidence
PixlNext, Prototyp3, CUSECCommunication and leadership
Explaining decisions and guiding teams toward a result.
One decision, in detail.
From my Fractal internship: shaping one architecture across several AI agent platforms.
Step 1
Compare
Built contextualization agents on several AI agent platforms, and kept what each one revealed.
Step 2
Test
Evaluated output quality, tool extensibility and platform constraints.
Step 3
Standardize
Defined source-backed guardrails and standard inputs, outputs and test cases so workflows repeat.
Have something in mind? Let’s talk.