Building scalable cloud systems and
practical AI applications.
I am Sunehildeep Singh, a Lead Software Engineer based in Toronto. I specialize in distributed Azure cloud infrastructure, high-throughput FastAPI backends, and responsive web applications. With over 3 years of production experience, I focus on building reliable systems that scale smoothly.
How I Think and Build.
End-to-End Ownership
I take full responsibility for the entire product lifecycle. That includes setting up cloud infrastructure on Azure, designing database models, building fast APIs, and polishing the frontend interface. If something breaks in production, I take ownership and resolve it quickly.
Reliability and Performance
I design backends to handle heavy user traffic without slowing down. By using asynchronous message queues like RabbitMQ and caching layers like Redis, I ensure systems stay responsive and databases remain protected under load.
Practical AI Solutions
I integrate machine learning and modern LLMs only where they solve real business problems. Whether it is automating document parsing or classifying high-volume orders, the solution has to provide clear value and fast response times.
Selected Projects &
Case Studies.
A collection of production cloud backends, AI research, and open-source software built for scale.
VeloCV: AI Resume Tailoring SaaS
A production SaaS that automatically tailors resumes to match specific job descriptions using NLP and LLMs. I redesigned the backend using stage-isolated RabbitMQ worker queues, which brought end-to-end generation time down to 20 seconds for over 800 active users on lightweight infrastructure.
- Built asynchronous RabbitMQ worker queues, preventing memory bottlenecks during high multi-user traffic.
- Engineered automated document generation pipelines supporting both LaTeX and clean PDF exports.
- Designed the full Next.js and Tailwind CSS frontend with a clean, fast onboarding experience.
Towards Inquisitive AI (Whitepaper)
A published research paper exploring how language models can autonomously identify gaps in their knowledge and ask clarifying questions.
Custom Fan Helios
An open-source hardware utility that lets users create custom fan curves using low-level Windows APIs. Built to replace bloated OEM software and currently used by over 500 active users.
Transformer AI Model from Scratch
An encoder-decoder sequence model built from scratch using Python and TensorFlow. Implemented multi-head self-attention mechanisms and positional embeddings based on the original Attention Is All You Need paper.
LSTM Character-Level Text Generator
A character-level recurrent neural network with GRU cells that generates dynamic text with adjustable temperature sampling.
SA:MP Multiplayer Game Systems
Custom multiplayer server plugins and netcode built in Pawn and C++. Handled real-time state synchronization, anti-cheat detection, and low-latency gameplay for over 1,000 active players.
IELTS Exam Preparation Platform
An automated grading tool that evaluated student writing and listening tests, saving instructors 70% of their manual grading time during remote learning.
Work, Leadership &
Education.
Track record of progression from software engineer to technical lead and cloud architect.
Lead Software Engineer
- Own overall technical strategy, architecture design, and code review standards while mentoring engineers to deliver features independently.
- Architected the complete migration of company infrastructure to Azure (AKS, App Service, Blob Storage, Azure SQL, and Relay Connections) with automated CI/CD pipelines.
- Engineered a cross-platform field operations app using Expo and React Native (iOS, Android, Web) featuring live location tracking, AI route optimization, and dynamic form generators.
- Resolved database timeouts in order management by adding a Redis caching layer over Azure Relay connections to on-premises databases.
- Built containerized FastAPI microservices and RabbitMQ message queues on Kubernetes for reliable multi-environment deployments.
Software Engineer
- Built a high-performance FastAPI rules engine to validate product constraints, dimensional formulas, and component compatibility across 120,000+ monthly B2B orders.
- Led the migration of a legacy PHP portal to Next.js and TypeScript, reducing weekly inbound customer support tickets by approximately 40%.
- Built a machine learning classification pipeline using scikit-learn to automate order routing, cutting manual data entry by 50%.
- Implemented Vitest testing framework achieving 85% code coverage and configured automated GitHub Actions CI/CD pipelines (reducing deploy times from 1 hour to 10 minutes).
AI Engineer (Contract)
- Built an end-to-end NLP and dual-LLM (GPT-4) pipeline for an AI matchmaking product to extract sentiment and align shared user interests.
- Implemented VADER sentiment analysis and TF-IDF cosine-similarity algorithms to match user profiles effectively.
- Developed geospatial midpoint filtering using the Google Maps API to recommend venues that match dietary and cultural preferences.
Software Engineering Technology: Artificial Intelligence
- Specialized in deep learning architectures, neural networks, and computer vision systems.
- Comprehensive coursework in software design patterns, distributed databases, and cloud engineering.
- Graduated with a 4.3 / 4.5 GPA across all academic terms.
Google Foobar Challenge Invitee
2018Invited to Google's selective developer challenge based on algorithmic problem-solving patterns.
GitHub Arctic Code Vault Contributor
2020Open-source repositories selected and archived in the Arctic World Archive to preserve code for future generations.
Published Paper: Towards Inquisitive AI
2024Authored research paper on autonomous continuous-learning LLMs achieving 30% inference improvements on benchmark datasets.
Technical
Proficiencies.
Full-stack depth across machine learning, scalable cloud backends, and modern interface development.
AI & Neural Systems
Cloud & Infrastructure
Frontend & Mobile
Languages & Backend
Databases & Storage
Reliability & Methods
The Journey.
From writing multiplayer game server scripts at age 14 to managing enterprise cloud infrastructure and publishing AI research.
First Lines of Code (Age 14)
Started programming by writing multiplayer game server scripts in Pawn and PHP, learning the fundamentals of raw memory management and network synchronization.
Honor Gaming Studios & Machine Learning
Co-founded an independent gaming studio, managing community members while building neural network models in TensorFlow and system automation tools in Python and C#.
Google Foobar & Arctic Code Vault
Invited to Google's selective Foobar developer challenge and contributed open-source code preserved in the GitHub Arctic Code Vault for long-term preservation.
Centennial College AI Specialization
Completed the 3-Year Software Engineering Technology: AI program with a 4.3 / 4.5 GPA, mastering deep learning architectures and distributed backend systems.
Custom Fan Helios Open Source
Released an open-source hardware utility that gained over 500 active users by communicating directly with low-level Windows APIs.
Software Engineer & Published Paper
Joined Sun Glow to build high-throughput FastAPI validation engines and modern Next.js portals. Published the Towards Inquisitive AI research paper.
Lead Software Engineer
Promoted to Lead Software Engineer at Sun Glow, managing Azure cloud infrastructure, Kubernetes microservices, and cross-platform mobile apps.