Company Description KIDS' SHIELD SERVICES Inc. (KIDS SHIELD Co.) develops cutting-edge cybersecurity programs and services to protect children from cybercriminals. The organization focuses on smart, safe, and responsible solutions that allow children to learn, connect, and explore online with confidence. By combining advanced intelligence and creative approaches, KIDS SHIELD Co. designs tools that help families and educators manage digital risks.

The company’s mission-driven culture emphasizes impact, ethics, and continuous innovation in child-focused online safety.

Role Description The Machine Learning Specialist will design, build, and optimize machine learning models that enhance child online safety and detect cyber threats. Day-to-day responsibilities include developing algorithms for threat detection, analyzing large-scale data sets, experimenting with deep learning techniques, and validating models for accuracy and reliability. The role also involves collaborating with cybersecurity experts, product teams, and engineers to integrate ML solutions into production systems, as well as documenting methodologies and presenting findings to stakeholders.

This is a full-time hybrid role based in Vancouver, BC, with a mix of on-site collaboration and work-from-home flexibility.

Qualifications

- Strong foundation in Computer Science, including data structures, software development principles, and system design.
- Expertise in Machine Learning and Deep Learning, with experience training, evaluating, and deploying models in real-world applications.
- Solid knowledge of Statistics and Algorithms to design robust, efficient,



and explainable ML solutions.
- Proficiency in relevant programming languages and tools (e.g., Python, TensorFlow, PyTorch, scikit-learn, SQL or NoSQL databases).
- Experience working with large, complex datasets, data preprocessing, feature engineering, and model performance optimization.
- Ability to communicate technical concepts clearly to both technical and non-technical stakeholders and collaborate in cross-functional teams.
- Awareness of cybersecurity concepts and interest in child online safety; prior experience in security or safety-focused products is an asset.

Key responsibilities· Audit the existing codebase, models, data flow and infrastructure; recommend keep/refactor/archive/replace. · Design a modular, testable architecture for interaction-risk intelligence rather than an end-to-end moderation platform.

· Define an initial child-safety risk taxonomy with the founder and market-validation specialist.

· Build a reproducible evaluation harness and establish strong baselines using general-purpose LLM prompting and generic moderation approaches where appropriate.

· Prototype temporal/sliding-window analysis, interaction memory and risk-trajectory scoring.

· Investigate game-language/domain adaptation, including slang,



coded language, evasion and context-dependent competitive speech.

· Design the dataset strategy: lawful/ethical sources, synthetic augmentation where appropriate, labeling protocol, train/validation/test separation and privacy controls.

· Measure precision, recall, F1, false positives/negatives, early-detection lead time, calibration, latency, throughput and inference cost.

· Evaluate multimodal expansion only after the text/interaction baseline is understood or when customer evidence establishes voice/image as the priority.

· Document experiments, assumptions, limitations and reproducibility.

Required skills· Strong Python and applied ML/NLP experience.

· Hands-on experience with transformer/LLM systems, embeddings/classification, evaluation and model/API integration.

· Experience designing rigorous benchmarks rather than relying on demo examples.

· Understanding of latency, throughput, inference cost and production ML tradeoffs.

· Comfort refactoring experimental ML code into modular, testable components.

· Ability to communicate technical uncertainty clearly and avoid unsupported performance claims.

Highly desirable· Trust & safety, online harms, child safety, gaming, cybersecurity or abuse-detection experience.

· Temporal/sequential modeling, conversational analysis or behavioral analytics.

· Model optimization, distillation, quantization or efficient inference.

· Speech/ASR or multimodal ML experience for later phases.

· Privacy-aware ML and experience handling sensitive datasets.

📌 Machine Learning Specialist (Vancouver)
🏢 KIDS' SHIELD SERVICES
📍 Vancouver