InfoQ has opened enrollment for the InfoQ Certified AI-Assisted Engineering Program, a new online certification program facilitated by senior practitioners in this space, Zichuan Xiong and Premanand Chandrasekaran. The first cohort runs September 18 through October 16, 2026, on Fridays at 9:00 AM PDT, for $1,470. A second cohort runs October 19 through November 16, Mondays at 9:00 AM PDT.
This certification is for senior engineers and architects who already run a coding agent on production code daily, working on codebases with years of history rather than simple AI demo applications. The program covers what an AI agent is allowed to do, how to catch it when it's wrong, and how you'd prove to your own organization that the code created is valid, scalable, and maintainable.
Xiong is Head of AIOps at Thoughtworks, has led architecture and delivery work since 2008, and builds agentic AI systems for software operations. Chandrasekaran is Head of Technology at Thoughtworks and has led engineering teams for two decades across financial services, retail, education, and healthcare. Both will run the Effective AI-Native Development with Harness Engineering training at QCon San Francisco on November 20.
Xiong says, "Teams are adopting coding agents faster than they're building the verification around them. The harness is everything outside the model: permissions, sensors, review gates, and CI checks that let an agent correct itself before a human ever sees the change. Over five weeks, you build a harness against a real codebase, and senior engineers from other companies tell you where yours is thin."
What each week covers #
Each week, participants watch a QCon talk in advance, then spend four hours in a live session applying that week's framework to a shared brownfield codebase alongside senior engineers from other companies.
- Week 1: onboarding an agent to an unfamiliar codebase, with least-privilege permissions and sandboxing.
- Week 2: turning a poorly specified issue ticket into a requirement an agent and a reviewer can verify against.
- Week 3: separating generation from review, using an independent harness rather than letting the generating agent grade its own work.
- Week 4: moving verification into CI, including governance for unattended agent runs.
- Week 5: codifying review findings into reusable rules, plus capstone presentations comparing five weeks of logged results against week-one predictions.
Live sessions and certification #
The program includes 20 hours of live sessions over five weeks working on the harness each participant is building. Sessions are not recorded, so participants can discuss permissions models, review gates and failures from their own organizations. The cohort and both facilitators continue in a private Slack channel between sessions.
Certification requires consistent attendance and a capstone written submission. Participants work a single brownfield repository across all five weeks, then submit the harnessed repository, one rule or agent skill codified from their own recurring review findings, and a write-up comparing their logged results against their week-one predictions.
The InfoQ AI programs #
AI-Assisted Engineering runs alongside AI Security & Privacy Engineering, with Katharine Jarmul. This program covers securing and governing AI systems in production: sensitive data handling, threat modeling and red teaming, controls and sandboxes, observability and evals, governance and auditing. This program begins August 26, and runs each Wednesday at 2:00 PM CEST.
Two more cohorts outside AI will be available soon: Architecture with Luca Mezzalira, starting September 14 (Mondays, 9:30 AM BST), and Engineering Leadership with Michelle Brush, starting September 18 (Fridays, 10:00 AM PDT). All five cohorts run for five weeks, four hours a week, and cost $1,470.
See all InfoQ Online Certification Programs at certification.qconferences.com.