Next.js · TypeScript · Node.js · MongoDB · Redis · OpenAI
Almaari is an AI-assisted digital wardrobe that combines clothing organization, outfit creation, and intelligent styling in one product.
View Case StudiesFull-Stack Developer · AI-Powered Products
I build useful software from idea to production.
I’m Manveer Sohal, a full-stack developer focused on AI-assisted products, backend systems, and polished web experiences. I created Almaari and JobLinx and have worked across generative AI, data systems, and computer vision.
Based in Ontario, Canada · Open to full-stack, backend, and AI-focused software development opportunities
Featured Projects
Projects I'm proud of
Projects I built to solve real problems I faced. Each with a case study.
Next.js · TypeScript · FastAPI · Supabase · TanStack Query · OpenAI
An AI-assisted job search workspace for organizing applications, analyzing job postings, tailoring application materials, and preparing outreach.
View Case Study
Next.js · TypeScript · FastAPI · pgvector · OpenAI · Supabase
A multi-tenant support copilot that finds relevant company documents, drafts grounded replies, and escalates when confidence is low.
View Case StudyExperience
Product building in practice
Hands-on ownership across product engineering — not just feature tickets.
2024–Present
Present
2024–Present
Present
Creator and Full-Stack Developer
AlmaariNext.js, TypeScript, Node.js, MongoDB, Redis, AWS, Python, OpenAI, Auth0, Docker
Designed and shipped an AI-assisted wardrobe platform end to end — from product flows to deployment.
- Designed and developed a responsive full-stack application using React, Next.js, TypeScript, Node.js, and FastAPI, implementing reusable components, authenticated user workflows, and RESTful API integrations.
- Improved application scalability and responsiveness by implementing Redis caching, database indexing, AWS S3 image storage, and CloudFront content delivery, reducing query latency by 50%.
- Developed asynchronous FastAPI services for AI-powered image analysis and metadata generation, automating onboarding workflows and reducing manual form completion by 40%.
2025
Geese Hacks
2025
Geese Hacks
Founding Software Engineer
Odin AnalyticaYOLOv8, Python, FastAPI, Mapbox, Computer vision
Early-stage traffic analytics venture that began as a Geese Hacks project and grew into a computer-vision product.
- Built a real-time traffic analytics system using computer vision and web dashboards
- Developed Python and FastAPI services for camera ingestion, validation, and ML inference
- Designed an asynchronous publish-subscribe architecture for dashboards and analytics processing
- Reduced processing latency by approximately 20%
2023
May – Dec
2023
May – Dec
Front-End Web Developer
Almaari DecorReact, Tailwind CSS
Built and shipped responsive React and Tailwind interfaces for a local business, connecting stakeholder requirements to polished web workflows.
- Built and shipped maintainable React and Tailwind CSS user interfaces using reusable components and a mobile-first approach, delivering responsive experiences across mobile, tablet, and desktop devices.
- Collaborated with stakeholders to translate business requirements and user feedback into intuitive web workflows, improving usability and increasing average session duration to 2.8 minutes.
- Developed end-to-end contact and inquiry workflows by integrating frontend forms with backend services and business processes, increasing consultations by 35% and reducing response times by 50%.
Technical skills
Tools I use to ship
Frontend
TypeScriptJavaScriptReactNext.jsHTMLCSSTailwind CSSTanStack Query
Backend
PythonFastAPINode.jsExpressREST APIsSQLAlchemy
Data & infrastructure
PostgreSQLSupabaseMongoDBRedisDockerAWSS3CloudFrontCI/CD
AI & machine learning
OpenAI APIsRetrieval-Augmented GenerationEmbeddingspgvectorComputer visionOpenCVYOLOPrompt & structured-output design
Additional work
Other projects
Earlier and smaller builds that still show range across realtime systems, analytics, games, and ML experiments.
Demo unavailable
Creator Control Room
A realtime analytics dashboard for tracking Twitch creator events such as gifts, followers, and subscriptions.
TypeScriptReactNext.jsNode.js
Live product
Clash Royale Hub
A Clash Royale stats hub for player profiles, battle logs, clan views, war logs, and a searchable card browser.
ReactTypeScriptNode.jsRedis
Live product
NBA Shooting Dashboard
An interactive dashboard for analyzing NBA player shooting and season statistics.
PythonPandasMatplotlibSeaborn
Live product
Codegories
A realtime multiplayer Scattergories-style game with server-rendered rooms and Socket.io updates.
TypeScriptNext.jsNode.jsSocket.io
Hackathon project
StudySync
A Hack the North study companion that transcribes speech and video, summarizes content, and generates quizzes.
OpenAIAuth0
Archived demo
Moodify
Predicts emotional state from recent Spotify listening using K-Nearest Neighbors on audio features.
PythonPandasscikit-learnSpotify API
Recognition & Milestones
Wins and presentations from hackathons and communities over the past few years.
- odinAnalytica — GeeseHacks Winner: Best Data Hack — crash detection on live traffic feeds with computer vision.
- StudySync — Hack the North Winner: Study companion that turns speech and video into summaries and quizzes.
- FairHold — ETHGlobal New York Winner: Tap-to-pay vendor deposits with milestone escrow, fewer disputes, and optional yield.
- Age:AI — Presented at Age:AI: Presented product and AI work at Age:AI in Waterloo.
Get in touch
Send a message
Tell me a bit about the role, collaboration, or product idea. Prefer email directly? manveersohalwork@gmail.com