Youssuf Helaly

Software Engineering Student at the University of Ottawa · 3.95 GPA

SoftwareEngineer

Five roles across design software, RF intelligence, cybersecurity, telecom, and video security. I ship AI agents, ML pipelines, cloud infrastructure, and backend services that run in production, currently at Autodesk.

Cloud Engineering
AI / Machine Learning
Backend Systems

View Resume Get in touch

Experience at

01/Experience

Five companies. Production impact at each.

Not a list of responsibilities. These are the systems I designed, built, and shipped, and what they changed.

Autodesk

Software Engineering Intern

2026 to Present

Building an AI agent that diagnoses issues and automates administration for Autodesk's enterprise platform.

Designed and built an end-to-end AI agent using LLMs and Model Context Protocol (MCP) that diagnoses issues and automates administrative tasks through natural-language, multi-turn workflows.
Engineered Java backend services and secure MCP tools with resource-level authorization, tenant isolation, and deterministic diagnostics for reliable AI-driven administration.
Analyzed 204 support escalations with Product, Support, and Engineering to identify and automate high-impact workflows for a channel handling 50 issues daily.

Java
MCP
LLMs
AI Agents

ThinkRF

Cloud & ML Engineering Consultant

2026

Architected the machine learning platform behind a global radio frequency spectrum monitoring network.

Architected a multi-model ML platform that processes 50M+ real-time RF observations from 50+ sensors worldwide, combining Isolation Forest and LSTM models for spectrum anomaly detection and classification.
Engineered a high-throughput AWS data pipeline ingesting 10M+ records daily, with automated CGI enrichment mapping cellular identities across 5 countries in real time.
Built an AI spectrum intelligence engine on Claude with structured prompt chaining that turns raw RF telemetry into root-cause reports, replacing 3+ hours of manual analysis every day.

Python
AWS
LSTM
Isolation Forest
Claude API

Trend Micro

Software Engineering Intern

2025

Cybersecurity simulation at scale on a Go-based digital twin platform at a global security company.

Led the design of an AI world generation system using MCP that autonomously creates 100+ realistic environments with dynamic network topologies and embedded CVEs, cutting manual setup by 80%.
Independently built a real-time security event processor that converts raw logs into structured attack narratives with phase detection and impact scoring, handling 1,000+ events per attack.
Extended the digital twin platform with network and API improvements plus Prometheus monitoring, enabling 30% more complex simulations.

Go
MCP
Prometheus
LLMs
Cybersecurity

Nokia

Full-Stack Engineering Intern

2025

Reliability and performance work across the stack of enterprise telecom software.

Raised system reliability by 20% by fixing service-layer defects and eliminating recurring Tomcat errors in Java Spring.
Automated Elasticsearch index configuration, cutting data inconsistencies by 40% and raising backend data quality.
Rebuilt React workflows with Redux and tighter API integration, reducing page load times by roughly 25%.
Authored PostgreSQL migration scripts that preserved 100% data integrity while accelerating database update cycles.

Java
Spring
React
Redux
Elasticsearch
PostgreSQL

Solink

Software Engineering Intern

2024

API development and test automation for a cloud video security platform.

Developed Python RESTful APIs managing 20+ YAML-configured endpoints, enabling rapid feature launches alongside frontend and product teams.
Designed 50+ automated test cases and automated their execution, cutting manual testing by 30% and saving $2,000+ per month.
Diagnosed and fixed CI/CD failures and flaky tests, reducing regression false positives across releases.

Python
Pytest
Playwright
REST APIs
CI/CD

University of Ottawa

B.A.Sc. Software Engineering, 2023 to 2027

3.95 / 4.0

Cumulative GPA

Data Structures and Algorithms, Operating Systems, Computer Architecture, Software Quality Assurance

02/Projects

Products, not assignments.

Side projects I designed and built end to end, from ML pipelines to the backends that serve them.

Snug

Live demo Source

See if real furniture fits your room before you buy it

Renters buy furniture from photos on a white background and find out it doesn't fit when it arrives. Snug scans your room on any iPhone, drops real Amazon products in at true size, and tells you honestly whether each one fits.

Launched to 25 beta users: AR room scanning on any iPhone, no LiDAR, measured at a 3.4 cm median wall error against a tape measure.
Trained a custom 11-class YOLO furniture detector (0.74 mAP50) on merged open datasets and deployed it on-device with CoreML, so the app spots existing furniture while you scan.
Fit-check engine with separating-axis collision, non-convex rooms, and a ±5 cm uncertainty band that says "too close to call" instead of faking precision, covered by 197 unit tests in CI.
Mesh-simplification pipeline that cut bundled 3D assets by 96% (682 MB to 28 MB), bringing the app under the App Store's 200 MB cellular limit, plus a TypeScript port of the fit engine for a live web demo.

Swift
SwiftUI
ARKit
RealityKit
CoreML
YOLO
TypeScript
three.js

ClipIt

Private build

Real-time ML video intelligence, entirely on device

Finding the moment that matters in long footage means scrubbing manually. ClipIt detects highlight-worthy moments as they happen, on the phone, with no cloud round trip.

Three-tier ML pipeline built on YAMNet and contextual models, reaching sub-500ms latency at 85% detection confidence on live video.
On-device inference with TensorFlow.js and WebGL acceleration over a rolling video buffer built on memory-efficient TypedArrays.
Rust backend on Actix-Web and FFmpeg that stitches buffered segments into clips and extracts audio for classification.

React Native
TypeScript
Rust
TensorFlow
Firebase

Image Analysis web interface with drag-and-drop upload and object search

Image Analysis Platform

Source

Computer vision search over thousands of images

Bulk image libraries are unsearchable without tags. This platform runs AWS Rekognition over every upload and makes whole collections queryable by object in seconds.

React and Flask application backed by AWS Rekognition and S3, improving search efficiency by 40%.
Handles authentication, bulk uploads, and object-count filtering, processing 500+ images in seconds.
Optimized SQLite queries to cut processing redundancy and keep response times flat as collections grow.

React
Flask
AWS Rekognition
S3
SQLite

More projects

Event Attendance Manager

Android app with role-based access for attendees, organizers, and administrators, backed by Firebase real-time sync.

Java
Android
Firebase
SQLite

Friend Recommender

Graph-based recommendation engine built on adjacency lists and common-connection scoring across social network data.

Python
Graph Algorithms

Slice Storm

2D arcade slicing game with dynamic difficulty scaling, full pause systems, and object-oriented architecture.

Python
Pygame

This Portfolio

The site you are reading. Next.js App Router, Tailwind CSS v4, server components, and structured data for SEO.

Next.js
TypeScript
Tailwind CSS

03/Skills

Depth where it counts.

Every group below is backed by production work, not a tutorial. The line under each one says where it was earned.

Cloud & Infrastructure

10M+ record daily pipelines at ThinkRF; simulation infrastructure at Trend Micro.

AWS
Docker
Kubernetes
Linux
Prometheus
CI/CD

Backend Engineering

Java services at Autodesk, Go services at Trend Micro, Java Spring at Nokia, Python APIs at Solink, Rust on side projects.

Go
Java
Spring Boot
Python
Flask
Rust
REST APIs

AI & Machine Learning

MCP-based AI agent at Autodesk; anomaly detection models at ThinkRF; LLM-driven world generation at Trend Micro; a custom YOLO detector on-device in Snug.

TensorFlow
PyTorch
YOLO
LSTM
Isolation Forest
Claude API
MCP
On-device inference

Frontend Development

Enterprise React workflows at Nokia; cross-platform mobile with React Native.

React
Next.js
TypeScript
Redux
React Native
Tailwind CSS

iOS & Spatial Computing

Snug: AR room scanning on any iPhone, a tested fit-check engine, and on-device CoreML detection.

Swift
SwiftUI
ARKit
RealityKit
CoreML
SwiftData
three.js

Databases & Data Engineering

PostgreSQL migrations with 100% integrity at Nokia; Elasticsearch automation in production.

PostgreSQL
Elasticsearch
Redis
SQLite
Firebase
SQL

04/Testimonials

What the people I worked for say.

Written by the senior engineers and managers who hired and mentored me.

He demonstrated excellent logical reasoning and a strong ability to understand and translate requirements into practical solutions. He was also capable of taking ownership of development tasks and delivering them with a good understanding of the software development lifecycle. His combination of technical competence and commitment allowed him to make a valuable contribution even in a relatively short period of time.

Tarcisio Franco de Carvalho

AWS Cloud Architect · ThinkRF

Despite being only in his second year of university, Youssuf consistently demonstrated a level of maturity and professionalism well beyond his years. He showed initiative in every aspect of his work and didn't hesitate to dive into unfamiliar territory to deliver results. It's rare to come across someone so early in their career who is this intentional about their development.

Sajjad Sherazi

Senior Full Stack Developer, Scrum Master · Nokia

Youssuf has been an invaluable intern with our team at Trend Micro. He played a crucial role in successfully completing a variety of tasks, including several that were highly complex. His ability to learn quickly and his eagerness to embrace new technologies were remarkable.

Munish Suri

Staff Software Developer · Trend Micro

He quickly adapted to new challenges, learning fast and making a meaningful impact on our automation efforts. His ability to take ownership and collaborate effectively made him a fantastic team player. I would definitely hire him again.

Cecilia Leong

Quality Architect · Solink

Full recommendations available on LinkedIn

05/About

How I got here.

I started shipping production code in my first year at the University of Ottawa and have not stopped since: test automation at Solink, full-stack telecom software at Nokia, cybersecurity simulation at Trend Micro, the ML platform behind a global RF sensor network at ThinkRF, and now AI agents for enterprise administration at Autodesk.

I care about systems that survive contact with real load. Pipelines that ingest millions of records a day. Models that run on devices instead of in demos. Code that other engineers can build on without reading my mind.

I am at my best when handed a hard, ambiguous problem and the ownership to solve it end to end.

Ottawa, Canada
B.A.Sc. Software Engineering, 2027
3.95 / 4.0 GPA
Currently interning at Autodesk

Contact

Let's build something that matters.

I am open to internship and early-career opportunities in cloud, AI/ML, and backend engineering. Email me and I will reply quickly.

youssufhelaly@gmail.com View Resume