Available for Co-op · Sept 2026

Khush Mittal

CS + Finance

Building at the intersection of software engineering and financial systems.

Resume
GitHub
LinkedIn
Email

About

Hello, I'm Khush Mittal.

I'm a CS + Finance student at the University of Waterloo, currently interning as a Web Developer at Mitsubishi Heavy Industries (Aerospace). I'm pursuing a Bachelor of Computing and Financial Management, minoring in Economics, and I'm a CFA Level I Candidate.

My interest lies at the intersection of software engineering, applied machine learning, and financial systems. I build tools that combine LLM-driven reasoning with real financial data, including Pocket, an AI-powered personal finance assistant, and a distributed task scheduling engine for reliable backend infrastructure. I currently collaborate with the Waterloo IR Group on information retrieval and applied NLP research, reproducing BM25, dense, and learned-sparse retrieval baselines for the Anserini and Pyserini projects.

Outside of academics and work, I play drums and football.

Projects

2 projects

Pocket — AI-Powered Personal Finance Assistant

LLM-driven financial insight engine with live bank data and overdraft-risk detection

  • Designed a multi-stage LLM prompt pipeline injecting live financial context (balances, transactions, budgets) into Gemini to generate structured, schema-compliant, production-ready responses — including overdraft-risk warnings and spending insights.
  • Built a Go REST service with JWT auth, rate limiting, and versioned prompt templates; instrumented usage monitoring to track response quality KPIs across real user sessions.
  • Containerized PostgreSQL + Redis with Docker and shipped through a CI/CD pipeline with automated tests; maintained architecture docs and prompt template versioning mirroring production engineering standards.
GoGeminiLLM Prompt PipelinePostgreSQLRedisPlaid APIDocker

Distributed Task Scheduler

Production-grade job engine with priority queues, worker-pool concurrency, and at-least-once delivery

  • Engineered a distributed job scheduling engine with priority queues, worker-pool concurrency, and Redis-backed state for reliable at-least-once task execution under concurrent load.
  • Exposed a REST API for task submission, status tracking, and real-time KPI monitoring; containerized with Docker Compose for fully reproducible deployment.
  • Documented system architecture end-to-end — component diagrams, setup guide, environment config — mirroring production engineering standards.
GoPostgreSQLRedisDocker

Skills

AI & Generative AI

  • LLM Prompt Engineering
  • RAG
  • Gemini
  • Claude
  • ChatGPT
  • Microsoft Copilot
  • TensorFlow
  • scikit-learn

Languages & Web

  • Python
  • TypeScript
  • Go
  • SQL
  • HTML
  • CSS
  • C++
  • Java

Data & Analytics

  • Power BI
  • Tableau
  • Excel
  • Pandas
  • NumPy
  • Data Cleaning & Validation
  • KPI Tracking & Reporting

Backend & Tools

  • Node.js
  • REST APIs
  • PostgreSQL
  • Git
  • Docker
  • Google Cloud Platform
  • Automation Anywhere (RPA)
  • Jira

Experience

  1. Web Developer Intern — Technology & OperationsMitsubishi Heavy Industries (Aerospace)

    Apr 2026 – Present · Canada

    • Eliminated a recurring manual step for the operations team by designing an LLM-assisted workflow that auto-generated standardized, brand-compliant email templates and dispatched them via the existing mail system.
  2. Financial Analyst InternAgroFood Grains

    May 2025 – Aug 2025 · Remote

    • Reduced month-end reconciliation discrepancies by ~30% by designing validation controls to reconcile transaction-level data across 4 business units simultaneously.
  3. Audit InternGrant Thornton Ltd

    Jun 2024 – Sep 2024 · Remote

    • Reduced close-cycle errors by reviewing 100+ audit reports and processing 50+ monthly journal entries in Tally, surfacing variance summaries and documented exceptions that were formally adopted by the engagement team.
  4. Operations & Analytics InternKIA Motors

    May 2023 – Aug 2023 · India

    • Cut manual processing time during an ERP rollout by building 12 RPA automation rules in Automation Anywhere, eliminating manual handoffs across teams.

Education & Research

University of Waterloo

Bachelor of Computing and Financial Management · Minor in Economics

Expected Dec 2027
  • Velocity Innovation Challenge — Finalist: Designed a VR- and 5G-enabled banking concept selected as a top-5 finalist; pitched to Unity Technologies, BMO, and Rogers.

Research Collaborator · Information Retrieval & Applied NLP

Waterloo IR Group

2026 – Present
  • Contributed merged pull requests to Anserini and Pyserini by reproducing BM25, dense (BGE-base), and learned-sparse (SPLADE-v3) retrieval baselines on MS MARCO, broadening the group's retrieval reproducibility pipeline.

Credentials & Achievements

  • Velocity Innovation Challenge Finalist
  • CFA Level I Candidate
  • Microsoft Azure Fundamentals (AZ-900)
  • Bloomberg Market Concepts (BMC)