Machine Learning Engineer – Live Video Intelligence

About the Role

Trellion is building the real-time intelligence engine for the future of hiring. We analyze not just resumes and job posts, but live human performance through video, voice, and interaction dynamics. This role sits at the intersection of machine learning, real-time systems, and product engineering.

You will own the end-to-end ML stack behind live video interviews: design, build, deploy, and optimize systems that process video, audio, and behavioral signals in real time. This is a production, real-world inference pipeline with latency, reliability, and accuracy constraints.

Responsibilities

• Design and implement real-time video and audio processing pipelines
• Develop computer vision solutions for face, gaze, posture, and gesture analysis
• Implement speech recognition and paralinguistic signal extraction
• Fuse behavioral signals and build temporal models
• Build and operate live inference systems with strict latency budgets
• Deploy models for streaming workloads and optimize inference
• Extract features for interview intelligence scoring
• Integrate ML outputs into user-facing products

Requirements / What We Care About

• You can ship ML into production, not just train models
• You understand real-time constraints
• You think in signal quality, not just accuracy
• You can debug video, audio, and model pipelines under pressure
• You build systems that degrade gracefully
• You take responsibility for outcomes, not just code

Full Stack Required

Machine Learning & Signal Processing

• Python in production
• PyTorch or TensorFlow
• OpenCV, MediaPipe, or equivalent
• Audio signal processing
• Time-series and sequence models
• Multimodal learning fundamentals

Backend & APIs

• FastAPI or similar
• WebSockets or streaming APIs
• REST and event-driven systems
• Data pipelines and feature stores

Real-Time & Infrastructure

• Docker and containerized inference
• GPU deployment and optimization
• Low-latency systems
• Monitoring, logging, and model observability

Frontend Integration

• Familiarity with WebRTC pipelines
• Experience shipping real-time features to the browser
• Understanding how ML surfaces inside products

Compensation & Benefits

• Base salary: $150,000–$190,000 CAD
• Meaningful equity
• Hybrid work in Montreal
• Ownership over one of the core intelligence systems at Trellion
• High trust, low bureaucracy environment

How to Apply

Send your resume, GitHub, and any real-time ML systems you’ve built to [email protected]. If your experience is purely offline modeling, this role will not be a fit.