Yash BarveYash Barve

Yash H. Barve

Software Engineer

I'm a fifth-year Computer Science student at the University of Waterloo and a Business student at Wilfrid Laurier University. Since starting my full-stack development journey in 2021, I have grown especially passionate about backend technologies and scalable systems. I am currently on my 4th work term working as a Software Engineer at Nokia.

As I continue to deepen my expertise in web development, I'm also exploring emerging fields like Web3 and Machine Learning. I'm eager to keep learning, collaborate with brilliant minds, and gain meaningful work experience along the way.

Beyond tech, I'm an avid reader and a huge Formula 1 fan - delusionally cheering for Scuderia Ferrari and Charles Leclerc every race weekend.
PROJECTS
Fit4Me – Personalized Fitness Tracking App
·
Full-stack Android fitness app (team project) built with Kotlin + Jetpack Compose and a Node.js/Express backend with PostgreSQL.
·
Delivers personalized workout planning and session tracking with a focus on a smooth, user-centric experience.
·
Includes real-time chat using Socket.IO alongside REST APIs for core app functionality.
·
Implements intelligent matchmaking based on user goals, availability, and experience level.
·
In active development with an emphasis on frontend polish and backend performance.
android
jetpack compose
kotlin
node.js
express.js
typescript
postgresql
prisma
rest apis
socket.io
jwt
websockets
Check it out:
YouTubeGitHub
Chattr - Minimal real-time group chat
·
Minimalist real-time group chat app built with React, TypeScript, Express, and Socket.IO.
·
Supports creating and joining public rooms via unique room codes and instant message delivery.
·
Includes system notifications and live presence updates for a responsive chat experience.
·
Designed as a lightweight, deployable MVP with clean UI and responsive interactions.
·
Built to extend into features like private rooms, message history, and moderation.
react.js
typescript
tailwindcss
socket.io
node.js
express.js
rest apis
websockets
Check it out:
LiveGitHub
Portolio - Your Online Photos Portfolio
·
Modern MERN-stack photo gallery app for uploading and showcasing images.
·
Implements direct client-to-ImageKit uploads for fast, scalable media handling.
·
Uses JWT-based authentication for secure access and user sessions.
·
Built a responsive gallery layout with uniform image sizing for a clean UI.
react.js
javascript
tailwindcss
node.js
express.js
mongodb
imagekit
jwt
Check it out:
GitHub
Deep Features for Interactive Segmentation
·
Interactive image segmentation tool built as a university capstone project using PyTorch and a pre-trained ResNet-50.
·
Extracts deep convolutional features to segment objects in complex images beyond traditional RGB-based approaches.
·
Provides a Python GUI for user-guided foreground/background selection to steer segmentation.
·
Implements max-flow/min-cut graph cut algorithms for robust, near real-time object segmentation.
·
Performs well in challenging scenarios with color ambiguity and noisy backgrounds.
python
pytorch
resnet-50
computer vision
graph cuts
max-flow min-cut
numpy
matplotlib
scikit-image
python gui
Check it out:
GitHub
Crop Yield Prediction & Risk Mitigation (BU425 Final Project)
·
Built a machine learning framework to predict crop yields and assess agricultural risk using real-world climate and farming inputs (rainfall, temperature, pesticide use, crop type).
·
Trained interpretable yield models using Random Forest and Lasso Regression for feature selection and prediction.
·
Engineered temporal features and implemented an LSTM model for sequential yield analysis.
·
Designed a dynamic risk scoring system to support resource allocation and deliver actionable insights for farmers.
·
Collaborated as part of a group project focused on sustainable, data-driven decision-making.
python
pandas
scikit-learn
random forest
lasso regression
feature engineering
tensorflow
lstm
rnn
data visualization
Check it out:
GitHub
ChatYHB - An AI powered chat bot
·
Built a web chat app using Google’s AI SDK with Gemini 1.5 powering backend responses.
·
Supports text-based prompts with clean, well-formatted outputs for a better reading experience.
·
Persists conversation history so users can revisit previous prompts and responses.
·
Includes a one-click option to clear all stored prompts and responses when needed.
react.js
javascript
tailwindcss
node.js
express.js
mongodb
google ai sdk
gemini 1.5
Check it out:
LiveGitHub
Scriptsync - Add transcriptions to your videos
·
Built an end-to-end captioning tool that generates subtitles from uploaded videos and lets users edit text, timing, and accuracy.
·
Provides caption styling controls (font size, text color, outline) to customize the final look.
·
Burns captions into the video using ffmpeg (via WebAssembly) for a fully client-side editing experience.
·
Supports exporting/downloading the finalized captioned video for sharing across platforms.
·
Uses AWS S3 for storage and AWS Transcribe for automatic speech-to-text.
next.js
typescript
tailwindcss
aws s3
aws transcribe
webassembly
ffmpeg
ffmpeg.wasm
Check it out:
YouTube
View All Projects
EXPERIENCE
Nokia
Software Engineering Intern · Sep — Dec '25
·
Built an LLM-based evaluation service (React, FastAPI, Celery, MySQL) with parallel async processing and a multi-step evaluation pipeline.
·
Shipped a user/admin settings service by designing React + MUI flows, exposing 10+ FastAPI CRUD endpoints, and modeling a scalable MySQL schema.
·
Led system design for 2 new services by defining API contracts, data models, and component/control-flow diagrams.
·
Deployed microservices to Kubernetes by authoring manifests and integrating CI/CD through Jenkins and Make.
·
Built a daily Kubernetes CronJob to automate auth-service backups, saving ~2–3 hours of manual work per cluster restart.
·
Nominated by manager for the 2025 Co-op Student of the Year award.
react
mui
javascript
python
fastapi
celery
mysql
docker
kubernetes
jenkins
make
ci/cd
UW Product Management Club
Engineering Director · Sep '25 — Present
·
Built an application submission platform for exec and event applications using Next.js, TypeScript, and TailwindCSS.
·
Designed a PostgreSQL database on Supabase to support application forms, applicant records, and reviewer workflows.
·
Implemented frontend hooks to fetch, transform, and format backend data for consistent UI rendering.
·
Developed an internal dashboard to render, filter, and label applicants to streamline review and decision-making.
next.js
react.js
typescript
tailwindcss
supabase
postgresql
AutoTrader.ca
Software Engineering Intern · Sep — Dec '24
·
Profiled and optimized frontend performance for a high-traffic marketplace, reducing Largest Contentful Paint (LCP) by 28% through JavaScript refactoring, async asset loading, and image compression.
·
Improved Core Web Vitals to enhance user experience and SEO performance across key pages.
·
Analyzed site structure and crawl data using Screaming Frog SEO Spider, identifying trends and high-impact issues to prioritize fixes.
·
Improved scalability for 15K+ dynamic URLs by refining server-side rendering, optimizing API responses, and removing redundant scripts
·
Automated site audit workflows using Python + REST APIs to streamline reporting and recurring checks.
javascript
web performance
core web vitals
seo
python
rest apis
data analysis
Petrovision International LLC
Software Development Intern · Jan — Apr '24
·
Built reusable React components (banners, images, layouts) for a website redesign to improve UI consistency across pages.
·
Integrated REST APIs to power dynamic content and supported SEO improvements during the rebuild.
·
Collaborated closely with senior developers in an agile workflow and contributed to ad-hoc data analysis tasks.
react.js
javascript
rest apis
git
seo
agile
Lazaridis School of Business and Economics
Data Insights Specialist Intern · Mar — May '23
·
Collected, cleaned, and analyzed marketing data using Microsoft Excel and Python to support campaign decisions.
·
Built Tableau dashboards to visualize performance, surface trends, and track KPIs for stakeholders.
·
Identified optimization opportunities that helped reduce marketing spend by 15% while maintaining performance.
microsoft excel
python
tableau
data cleaning
data analysis
data visualization
UW Data Science Club
Web Designer Executive · Oct '22 — Aug '23
·
Partnered with event organizers to gather requirements and translate them into clear, on-brand designs.
·
Designed promotional creatives for Instagram, LinkedIn, Facebook, and 𝕏 using Canva.
·
Contributed to early website design work in Figma, focusing on layout and visual consistency.
canva
figma
visual design
social media design
communication
collaboration
View Resume
TECHNICAL SKILLS & TOOLS
LANGUAGES
C
C++
Bash
Go
Python
HTML
CSS
Javascript
Typescript
SQL
Kotlin
CORE CS
Object Oriented Programming
Data Structures
Algorithms
Time & Space Complexity Analysis
Operating Systems
Concurrency
Relational Databases
User Interfaces
Software Design Patterns
Software Architecture Styles
FRONTEND
React.js
Next.js
TailwindCSS
Bootstrap
BACKEND
REST APIs
Node.js
Express.js
FastAPI (Python)
Hono
Websockets
Socket.io
Auth (JWT)
Zod
DATABASE
PostgreSQL
MongoDB
MySQL
Firebase
DEVOPS
Jenkins
Make
GitHub Actions
Railway
Vercel
Heroku
TOOLS
Prisma ORM
Git
Docker
Kubernetes
npm
WebAssembly
Jest
Jira
Tableau
Figma
CLOUD
AWS
AI & ML
OpenAI API
Gemini SDK
Codex
Claude Code
GitHub Copilot
Neural Nets
CNNs
RNNs
Computational Vision
Tensorflow
PyTorch
Numpy
Pandas
Matplotlib
OTHER
Agile Methodology
Scrum
Jetpack Compose (Android)
Retrofit (Android)
EDUCATION
Bachelor of Computer Science with Specialization in AI
University of Waterloo · 2021 — 2026
Courses completed:
Algebra
Calculus
Linear Algebra
Optimization
Combinatorics
Probability
Statistics
Functional Programming
Algorithm Design and Data Abstraction
Object Oriented Programming
Data Structures
Algorithms
Logic and Computation
Computer Organization and Design
Sequential Programming
Introduction to Machine Learning
Operating Systems
User-Interfaces
Computational Vision
Information Systems Management
Software Design and Architectures
Distributed Systems
Computer Networks
The Social Implications of Computing
Bachelor of Business Administration with Finance Concentration
Wilfrid Laurier University · 2021 — 2026
Courses completed:
Functional Areas of an Organization
Business Environment
Introductory Microeconomics
Introductory Macroeconomics
Financial Accounting
Business Law
Managerial Accounting
Organizational Behaviour
Human Resources Management
Marketing Management
Business Operations
Data Analytics
Strategic Management
Financial Management
Risk Management & Insurance
Options, Futures and Swaps
Personal Finance Management
Full Stack Web Development Cohort
100xDevs by Harkirat Singh · 2023 — 2024
Description:
This 8-month cohort transforms you from a 1x engineer to a 100x engineer. Some of the important concepts taught are MERN stack, Next.js, PostgreSQL, DevOps, WebRTC, and advanced web development. Additionally, there are several open-source projects to apply your knowledge.
Certificate of completion
Machine Learning Specialization
Deeplearning.ai by Andrew Ng · 2022
Description:
This 3-course specialization provides a broad introduction to modern machine learning, including supervised learning (multiple linear regression, logistic regression, neural networks, and decision trees), unsupervised learning (clustering, dimensionality reduction, recommender systems), and some of the best practices used in Silicon Valley for artificial intelligence and machine learning innovation (evaluating and tuning models, taking a data-centric approach to improving performance, and more.)
Certificate of completion
Deep Learning Specialization
Deeplearning.ai by Andrew Ng · 2022
Description:
This 3-course specialization provides a broad introduction to modern machine learning, including supervised learning (multiple linear regression, logistic regression, neural networks, and decision trees), unsupervised learning (clustering, dimensionality reduction, recommender systems), and some of the best practices used in Silicon Valley for artificial intelligence and machine learning innovation (evaluating and tuning models, taking a data-centric approach to improving performance, and more.)
Certificate of completion
WRITINGS
Course Notes Collection
Summary:
Over five years at the University of Waterloo and Wilfrid Laurier University, I’ve completed 50+ courses across Maths, Computer Science, and Business. For most of these, I developed detailed technical notes while preparing for exams. I have hosted these notes on GitHub and Notion as a resource for future students and a personal reference for my own work.
When Should AI Explain Itself?
Summary:
When AI operates as a supervised tool, human professionals retain full accountability and can demand explanations as needed. However, autonomous AI agents in high-stakes domains must be architecturally required to document their reasoning, serving as a substitute accountability mechanism and enabling error detection and correction. Note: This paper was written for CS492, a CS ethics course at the University of Waterloo.
Newcomb’s Paradox: The Wrong Choice I’d Still Make
Summary:
What does your decision-making style say about you? Newcomb's Paradox is a thought experiment involving a near-perfect Predictor and two boxes. You're forced to choose between two equally intuitive but conflicting strategies. This essay explores both options, and makes the case for the one I'd pick.
Outsourcing the Mind: Cognitive Friction and the Evolution of Human Intelligence
Summary:
Computer technology has progressively eliminated the cognitive friction required to learn - from books, to search engines, to AI that now delivers pre-synthesized answers on demand. While the efficiency gains are undeniable, the essay argues that friction was the very mechanism through which knowledge was deeply encoded, and that outsourcing it entirely risks turning users into passive validators rather than active thinkers. Note: This essay was written for CS492, a CS ethics course at the University of Waterloo.
Position Paper: Defense of Proprietary AI-Trading Algorithms Against Mandatory Disclosure
Summary:
This paper argues that organizations should be allowed to keep proprietary AI-designed trading algorithms private rather than disclose them publicly. Public disclosure could increase market instability by making trading behavior predictable and easier to exploit, while also reducing incentives for firms to innovate. Instead, the paper proposes confidential regulatory transparency, where algorithms are shared only with regulators, allowing oversight without sacrificing intellectual property protection. Note: This paper was written for CS492, a CS ethics course at the University of Waterloo.
View All Writings
HOBBIES & INTERESTS
READING

Mostly fiction with occasional non‑fiction. I’m trying to read more consistently, and I keep a running log on the Bookshelf page.