AI Engineer
← All organizations

Developer tools and AI context engineering

Unblocked

Unblocked builds a context engine that brings together source code, documentation, issues and team conversations so developers and AI agents can understand software and the decisions behind it. Engineers use its web and desktop apps, Slack and Teams integrations, and code review product to ask questions and review changes. Agents access the same engine through MCP, a CLI and an API, bringing organizational knowledge into planning, implementation, bug validation and CI workflows.

Founded in 2021, the Vancouver company is led by founder and CEO Dennis Pilarinos. Its engineering approach combines semantic search, knowledge-graph traversal and relational queries, then compresses retrieved information into task-relevant context with citations. An expertise graph uses code authorship and review activity, weighted over time, to help resolve conflicting accounts of how systems work. Unblocked also released engineering-social-graph, a simplified open-source implementation of that expertise model.

In April 2026, the company reported that tens of thousands of developers used its product experiences every day. By August 2026, it described its largest customer organizations as operating well over 100,000 repositories with tens of thousands of engineers. Unblocked raised a US$20 million Series A in 2025, led by B Capital with participation from Radical Ventures and Artisanal Ventures.

getunblocked.com ↗

3 talks

Newest first

3 speakers at AIE

Affiliations reflect their AIE appearances, not necessarily current employment.

Jeff NgEngineerPeter WerryFounding EngineerBrandon WaselnukDeveloper Relations

Start here

  1. Building agents is trivial now, context is the next frontier

    Start with Jeff Ng's talk to understand why MCP connectivity alone cannot reconcile conflicting information or control context costs.

    Jeff NgAI Engineer World's Fair 2026

  2. Mergeable by default: Building the context engine to save time and tokens

    Choose Peter Werry's recording for context-aware pull-request review, audience Q&A, and an extended hands-on workshop.

    Peter Werry · BrandonAI Engineer Europe 2026

  3. Stop babysitting your agents: building a context engine for mergeable code

    Watch Brandon Waselnuk's talk for an MCP-assisted Zendesk integration demonstration that puts context assembly into a concrete workflow.

    Brandon WaselnukAI Engineer Europe 2026

Messages from the stage

Incident history changes an agent's plan

Ng illustrates the risk of missing organizational context with an agent that recommends restoring a setting implicated in an earlier outage. His talk distinguishes the oversight available to local agents from the demands of unattended operation.

Retrieval needs more than a larger window

Werry examines targeted, personalized retrieval alongside satisfaction-of-search failures and answer-caching pitfalls. The discussion connects context selection to team conventions and prior engineering decisions.

Context requires access and freshness controls

Waselnuk describes dynamically assembled context that combines social-graph relevance, permission-aware access, conflict resolution, and freshness controls. Token-efficient research packets provide another design consideration beyond adding tool connections.

Affiliations reflect each recorded session, not necessarily current employment.

Company sources · checked 2026-08-27