Machine Learning Engineering Co-op - VCycene Inc.

Machine Learning Engineering Co-op

Sensors & Embedded | VCycene Inc. | Markham, ON | On-site | 8 Months | Fall 2026 Term

Location

On-site — 3600 Steeles Ave. E., Suite B112, Markham, ON L3R 9Z7 (this role cannot be done remotely)

Term

September 2026 – April 2027 (8-month co-op / internship placement)

Compensation

$20–$25 CAD/hour, based on candidate qualifications

Reports To

Junaid Siddiqui, who leads ML & software development for LILA

Apply

huayi@virgohome.io

About VCycene

VCycene builds the LILA Mini and LILA Pro, countertop smart composters sold across Canada and the US, and the Lovely companion app that pairs with them. We're a small hardware team, which means the person who trains the model is usually the same person who ran the compost cycle that produced the data. A growing fleet of units in the field generates continuous sensor telemetry that drives the product's core features.

About the Role

You will join our ML team working directly on the models that make LILA work: predicting compost readiness, detecting cycle anomalies, and tuning the device's response to what is actually in the bin.

This is a full-loop role. You will run composting cycles on test units, instrument them, collect and label sensor data, train models against that data, and then evaluate how those models behave on real devices in customers' kitchens. Expect to spend real time at the bench with food waste — the work is hands-on and occasionally smells like it.

What You'll Do

· Run structured composting experiments on LILA test units, varying feedstock, load, and cycle conditions

· Instrument test rigs and collect time-series data from temperature, humidity, gas, weight, and motor-load sensors

· Build and maintain data pipelines that move device telemetry into our training datasets

· Clean, label, and version datasets; identify and correct sensor drift and labeling gaps

· Train, evaluate, and iterate on models for compost readiness estimation and cycle anomaly detection

· Validate model behaviour against fleet data from production devices and flag regressions

· Write internal analysis on model performance and experiment results

· Work with firmware and app engineers to ship model updates to devices and surface results in the Lovely app

Requirements

· Strong Python (pandas, numpy, scikit-learn or PyTorch)

· Familiarity with core ML concepts: training/validation splits, overfitting, evaluation metrics

· Comfortable getting hands dirty — this role involves handling food waste and compost daily

· Able to lift 25 kg (test units and feedstock)

· Experience working with sensors and sensor data, in coursework, research, or personal projects

· Familiar with REST APIs