Your New Company
A leading Canadian retailer at the forefront of food and health innovation, committed to delivering quality, value, and sustainability. With a diverse portfolio of trusted brands and a strong presence across grocery, pharmacy, and wellness, this organization plays a vital role in serving millions of Canadians every day. Their focus on community, environmental responsibility, and long-term growth makes them a cornerstone of the Canadian retail landscape.
Your Role:
In this role, you will bridge the gap between complex data orchestration and advanced machine learning services. You will be responsible for developing the "brain" of our agent—implementing multi-agent architectures, robust RAG pipelines, and safety guardrails—while ensuring our underlying data ecosystem (Vector DBs, SQL, and Knowledge Graphs) is perfectly tuned for high-stakes medical contexts.
Agentic Systems \& ML Services
Architect Multi-Agent Workflows:
Design and implement a 2-layer supervisor-router graph and subgraphs using
LangGraph
and
LangChain
to coordinate complex task execution.
Reasoning \& Planning:
Implement
ReAct patterns
to enable the AI to perform autonomous chain-of-thought reasoning, strategic planning, and tool-based actions.
Tool Integration:
Develop and maintain function calling capabilities and Model Context Protocol (MCP) servers to allow the agent to interact with external APIs and databases.
Safety \& Guardrails:
Build and deploy rigorous guardrail systems to detect and mitigate malicious inputs, handle medical crisis queries, and prevent inappropriate or biased outputs.
Evaluation Frameworks:
Build and maintain a comprehensive evaluation service to measure LLM performance, grounding accuracy, and agentic reliability.
Data Engineering \& Infrastructure
Data Acquisition:
Develop scalable web scrapers and data collection pipelines using
Scrapy
and
BeautifulSoup
.
Pipeline Orchestration:
Manage complex ETL/ELT workflows using
Apache Airflow
to process and ingest healthcare data.
Hybrid Data Storage:
Architect and optimize data ingestion into
Weaviate
(Vector DB) for semantic search and possibly future case for
Neo4j
(Knowledge Graph) for structured relationship mapping.
RAG Optimization:
Build and refine a full
Retrieval-Augmented Generation (RAG)
pipeline to ensure all LLM responses are grounded in verified healthcare data sources.
Minimum Qualifications
Advanced Python:
Expert-level proficiency in Python and its data ecosystem.
LLM Orchestration:
Proven experience with
LangChain
and
LangGraph
for building stateful, multi-agent systems.
Database Expertise:
Hands-on experience with Vector Databases (e.g.,
Weaviate
), Graph Databases (e.g.,
Neo4j
), and standard SQL.
Data Engineering:
Proficiency with
Apache Airflow
and web crawling frameworks (Scrapy/BeautifulSoup).
Experience \& Knowledge
RAG \& Grounding:
Deep understanding of embedding models, retrieval strategies, and grounding techniques to minimize hallucinations.
Agentic Patterns:
Practical experience implementing ReAct, Plan-and-Execute, or similar agentic reasoning patterns.
Safety \& Ethics:
Experience implementing LLM safety layers and handling sensitive user queries (preferably in a regulated domain like healthcare).
API Development:
Strong experience building and consuming RESTful APIs and implementing tool-calling interfaces.
Preferred Qualifications:
Experience with healthcare data standards (e.g., HIPAA compliance, FHIR).
Experience building and scaling production-grade evaluation suites for LLMs (e.g., PromptEval, RAGAS, LangSmith).
What You'll Get in Return
Competitive rate.
Challenging and great work environment