Applied Scientist

About ShyftLabs
At ShyftLabs, we live and breathe data. Since 2020, we’ve been helping Fortune 500 companies unlock growth with cutting-edge digital solutions that transform industries and create measurable business impact. We’re growing fast and are looking for passionate problem-solvers who are excited to turn advanced research into real-world, enterprise-scale AI systems.

The Opportunity
We are seeking a highly skilled Applied Scientist to work at the intersection of applied research and production AI, with a strong focus on knowledge-centric AI systems, graph-based learning, and advanced Retrieval-Augmented Generation (RAG) architectures.

In this role, you will design, prototype, and productionize state-of-the-art AI systems that combine LLMs, knowledge graphs, vector databases, graph neural networks, and multimodal models. You will play a key role in building next-generation AI platforms for Customer OS and Retail OS, powering intelligent search, reasoning, personalization, decision intelligence, and document understanding at scale.

This role requires a product-driven mindset, someone who can quickly understand complex business problems, frame them as AI challenges, and design efficient, scalable solutions that deliver measurable impact. You will collaborate closely with engineering, product, and business teams to translate ambiguity into robust AI systems deployed in real enterprise environments.

What You'll Be Doing:
• Conduct applied research to solve real-world problems using LLMs, graph-based models, and multimodal AI.
• Rapidly understand problem context, constraints, and success metrics, and design pragmatic AI solutions aligned with product and business goals.
• Design hybrid AI architectures combining knowledge graphs, vector search, and deep learning for reasoning-aware systems.
• Research and implement graph embeddings, graph attention networks (GATs), and graph neural networks (GNNs) for representation learning and inference.
• Design and build advanced RAG systems at scale, going beyond naïve vector similarity search.
• Implement hybrid semantic retrieval across vector stores and graph databases (e.g., entity-aware retrieval, path-based reasoning, graph-augmented RAG).
• Optimize retrieval pipelines for latency, relevance, grounding, and explainability in production environments.
• Fine-tune LLMs and embedding models for domain-specific tasks (instruction tuning, adapters, LoRA, etc.).
• Design and implement LLM agent systems, including multi-agent orchestration strategies, tool use, planning, and memory.
• Evaluate, iterate, and optimize agent architectures to solve complex, multi-step enterprise workflows efficiently.
• Build and fine-tune document extraction pipelines, including (OCR systems, Layout-aware models, Vision-Language Models (VLMs), Multimodal document understanding and classification)
• Design scalable pipelines for enterprise document ingestion, enrichment, indexing, and retrieval.
• Build end-to-end AI pipelines covering data ingestion, feature engineering, training, evaluation, deployment, and monitoring.
• Partner with platform and data engineering teams to productionize solutions on AWS or GCP.
• Monitor model performance, detect drift, and drive continuous improvement strategies.
• Design evaluation frameworks, offline metrics, and online experimentation (A/B testing) to measure real-world impact.

What You'll Bring:
• Bachelor’s, Master’s, or PhD in Computer Science, Machine Learning, Data Science, or a related field.
• Strong proficiency in Python and modern ML frameworks (PyTorch preferred).Hands-on experience with applied research and translating research ideas into production-grade AI systems.
• Proven experience with knowledge graphs, graph embeddings, or graph neural networks.
• Experience building advanced RAG systems using vector databases and structured knowledge sources.
• Strong understanding of LLMs, embeddings, and fine-tuning techniques.
• Experience deploying AI systems in enterprise or large-scale production environments.
• A product-oriented, problem-solving mindset with the ability to quickly learn new domains and design efficient AI solutions under real-world constraints.
• Solid foundation in ML fundamentals, statistics, and experimentation.

Nice to Have:
• Experience with graph databases (e.g., Neo4j, Neptune) and vector stores.
• Experience with agent-based LLM systems and multi-agent strategies.
• Experience with MLOps tools (SageMaker, Vertex AI, MLflow, feature stores).
• Familiarity with time-series forecasting, recommendation systems, or personalization models.
• Prior experience contributing to open research or collaborative academic projects

What Sets This Role Apart:
• Publications in top-tier conferences or reputed journals
• Strong research background with peer-reviewed papers
• Experience publishing in leading AI/ML venues (e.g., NeurIPS, ICML, ICLR, CVPR)
• Demonstrated research impact through high-quality academic publications

Salary Range:
• $110,000 - $150,000 (CAD)

Why You’ll Love Working at ShyftLabs
Hybrid Flexibility: 4 days per week in our downtown Toronto office.
Comprehensive Benefits: 100% coverage for health, dental, and vision insurance for you and your dependents from day one.
Real Impact: Build AI systems used by Fortune 500 companies.
Growth & Learning: Continuous learning opportunities and influence over technical direction.
Ownership: Shape applied research and AI strategy in a fast-growing, product-focused data company.

Inclusion at ShyftLabs
We’re building something big, and we want you on the journey with us. If you’re ready to use data and innovation to make an impact, apply today and let’s grow together.

ShyftLabs is an equal-opportunity employer committed to creating a safe, diverse, and inclusive environment. We encourage applicants of all backgrounds including ethnicity, religion, disability status, gender identity, sexual orientation, family status, age, and nationality to apply. If you require accommodation during the interview process, let us know and we’ll be happy to support you.