Overview
QCon AI Boston 2027 is a two-day conference for senior engineers focused on the problems that appear once AI systems run in production. It takes place July 8 to 9, 2027, in Boston. The program addresses long-running agents, company context that goes stale, evaluations that teams can act on, controlled tool access, and rising inference costs. The event is produced by C4Media, the team behind InfoQ and QCon, and the organizers state that sessions are led by senior practitioners who build and operate the systems they describe.
Program areas
The 2027 program is being built around six areas:
- Agent runtime design: state, recovery, and approvals for agents that run for minutes, hours, or days.
- Context engineering: keeping company context useful without flooding the prompt.
- Evals and observability: seeing what a system did and deciding whether a change ships.
- MCP and tool access: how much access an agent gets to internal systems and who approves it.
- Zero-trust agent security: identity, least privilege, and kill switches for agents that take action.
- Inference cost and latency: which requests need the most capable model and what each choice costs.
Sessions are not yet published as a full agenda. Speakers and sessions are added as they are confirmed.
Speakers
The following speakers have been announced so far:
- Cassandra Shum, Vice President of Field Engineering at RelationalAI
- Fabiane Nardon, Data Expert, Java Champion, and Data Platform Director at totvs, with more than 20 years of experience architecting large-scale data systems
- Ian Thomas, Software Engineer at Meta
- Meryem Arik, Co-Founder and CEO of Doubleword, previously at TitanML
- Shuman Ghosemajumder, Co-Founder and CEO of Reken and Co-Founder and Chairman of TeachAids
A Program Committee of working senior engineers invites speakers, and the organizers state that sponsors do not control the program. Sponsored content is kept separate from the talks.
Format
The conference runs for two days and includes sessions and keynotes, with one speaker for every eleven attendees. Hundreds of senior engineers attend, and the schedule leaves time to work through problems with peers and to ask speakers questions one on one. A ticket includes social events and receptions, continental breakfast, lunch, coffee breaks, and access to the exhibitor area. Buyers also get on-demand access to most presentation recordings and slides for 12 months after the conference.
Venue
The event is held at the George Sherman Union, 775 Commonwealth Avenue, Boston, MA 02215. QCon AI Boston is not affiliated with Boston University, and the use of its facilities does not imply endorsement. The organizers list discounted group rates at hotels near the venue through an official hotel portal.
Who should attend
The event is aimed at people responsible for how AI is designed, integrated, evaluated, secured, operated, or adopted in production software. This includes senior software engineers, software architects, Staff+ engineers, ML and MLOps engineers, platform engineers, security engineers, and engineering leaders. It is not an introductory AI conference. The content centers on architecture, reliability, security, evaluation, infrastructure, and cost.
Pricing and registration
A ticket costs $995 until November 10, 2026. The price then increases on set dates, reaching $1,695 from June 16, 2027. Groups of 10 or more receive $100 off each ticket until February 9, 2027. From February 10, groups of three or more receive $50 to $125 off per ticket, and QCon alumni receive $100 off. Payment is accepted by card or PayPal, with invoice and wire options on specified dates. Registration fees are non-refundable, but tickets can be transferred to a colleague for free until July 5, 2027.
FAQ
Is the program finalized? No. The organizers say speakers and sessions are added as they are confirmed.
Can I transfer my ticket? Yes, free transfers are available until July 5, 2027, by contacting the organizers.




