AI Engineer-Memory Retrieval

IDC Research Inc.•Toronto, ON

58d•Remote

About The Position

IDC is building the next generation of AI-powered intelligence platforms that transform how technology decisions get made. Our platform re-imagines the way decision-makers discover and interact with trusted research and data, and is foundational to IDC's future. We are looking for an AI/Agent Engineer to join the Memory / Living Intel Pod, with an emphasis on cross-conversational retrieval and semantic search over conversation history. This is a stretch and acceleration hire. You will build the retrieval systems that find relevant prior conversations and extracted facts, design embedding and indexing strategies for conversation-derived content, and optimize relevance ranking for memory retrieval.Your work ensures that when the platform remembers a user, it remembers the right things at the right time. The quality of memory retrieval directly determines how personalized and useful every AI-generated response feels

Requirements

3+ years of experience building production Python systems
Experience with LLM applications: agent orchestration, prompt engineering, RAG, or similar
Hands-on experience with agent frameworks (LangGraph, LangChain, CrewAI, AutoGen, or similar)
Understanding of LLM reasoning patterns and common failure modes
Proficiency with modern Python web frameworks
Comfort working in a fast-moving team where priorities evolve

Nice To Haves

Experience building semantic search and embedding pipelines
Hands-on experience with vector databases (e.g., Snowflake Cortex, Pinecone, Weaviate)
Background in information retrieval
Experience with search relevance tuning and ranking optimization
Familiarity with LLM observability tools
Background in NLP, text summarization, or information extraction

Responsibilities

Build specialized agents within multi-agent pipelines designed for handling complex research problems, developing durable outputs, and taking action in external systems on the user's behalf
Build retrieval systems that find relevant prior conversations and extracted facts across a user's history
Design embedding and indexing strategies for conversation-derived content
Build and optimize relevance ranking for memory retrieval
Build LLM-based memory capabilities: conversation summarization, cross-session context retrieval, and persistent user preference extraction
Instrument pipelines for observability, tracing, and quality monitoring
Collaborate with evaluation engineers on quality measurement and improvement