BT

Facilitating the Spread of Knowledge and Innovation in Professional Software Development

Write for InfoQ
Register Sign in

Unlock the full InfoQ experience

Unlock the full InfoQ experience by logging in! Stay updated with your favorite authors and topics, engage with content, and download exclusive resources.

Log In
or

Don't have an InfoQ account?

Register
News Articles Presentations Podcasts Guides

Topics

Choose your language

AI Security & Privacy Engineering Certification

Secure and govern production AI systems, from sensitive data to guardrails, evals, and audits.
Online. Register now.

AI Engineering Certification

Production AI calls on retrieval, agents, evals, and infrastructure, checked with peers.
Online. Register Now.

Architect Certification

Distributed systems, decentralized decisions, platform engineering, and AI architecture.
Online. Register Now.

QCon San Francisco

What's working across AI, architecture, and leadership, from the teams doing it.
Register. Early bird ends July 14.

QCon London

What early-adopter teams have proven in production, across 15 engineering tracks.
Register. Early bird ends July 14.
Recordings now available

QCon AI Boston
Solutions Track Library

Twelve production-focused talks from the engineers and platforms building today's agentic systems — recorded live at Boston University, curated and hosted by InfoQ.

Browse all recordings About QCon AI Boston

All solutions track recordings

Filter by day, track, or topic. Every recording is freely available, hosted on InfoQ.

13 talks
Solutions Track

From Prompt Hacking to Architectural Determinism: Engineering Reliable GenAI Systems

Prompt engineering can't deliver enterprise reliability—model updates, context drift, and window compression guarantee regressions. This talk introduces architectural determinism: a framework that makes the LLM a router for novel reasoning while delegating execution to deterministic, auditable code. The result is GenAI that's repeatable, governable, and safe at scale.

Hugo Guerrero
Kong
Watch →
Solutions Track

From Demo to Production: Why Agentic AI Systems Fail—and How to Fix Them

Most agentic AI systems look great in demos—and collapse in production. This talk reveals why: it's a systems problem, not a model problem. Through hard lessons building a platform for 10M+ users, learn how to master context management, tool exposure, agent boundaries, and observability to ship AI that actually works.

Venugopal Jidigam
WaveMaker
Watch →
Solutions Track

Leverage AI to Protect Your Data Assets... from AI

AI accelerates development—and data leaks. PII hides in nested NoSQL schemas, slips into vector stores, and surfaces in places no one intended. This talk shows how schema registries, automated PII scanning, and deterministic masking pipelines let you harness AI safely, by governing your data before AI governs it for you.

Daniyar Mussakulov
3T Software Labs
Watch →
Solutions Track

Solving Context Bloat: Semantic Tool Routing in Multi-Server MCP Environments

One MCP server. Five times the tokens. Two more servers? 3,500 tokens for a 20-token prompt. This talk shows how a gateway-based semantic routing layer solves context bloat by delivering only relevant tools at runtime—cutting cost, reducing hallucinations, and bringing battle-tested microservices patterns to the agentic era.

Hugo Guerrero
Kong
Watch →
Solutions Track

The Trust Lifecycle for Safe, Secure and Reliable AI Agents

AI agents fail in ways traditional software doesn't — biased outputs, goal hijacking, and silent compliance violations. This talk introduces the Agent Trust Lifecycle: a rigorous engineering framework covering reliability, safety, and security from development through production. Learn how testing at scale, runtime guardrails, and automated optimization keep agents reliably trustworthy.

Anuj Tambwekar
Vijil.ai
Watch →
Solutions Track

Vibecoding your Own Multi-Agent Workstation

No off-the-shelf AI coding tool fits every developer's workflow. This talk argues for building your own cloud-hosted, always-on multi-agent workstation — customized, event-driven, and fully shared across your team. Learn to fork an open-source foundation, start with low-risk review bots, and iterate toward agents that restore the genuine joy of programming.

Robert Brennan
OpenHands
Watch →
Solutions Track

Klaviyo Customer Agent: Building Bi-Directional Conversational Commerce at Scale

Marketing messages are one-way. Klaviyo's Customer Agent changes that — turning outreach into real-time, bi-directional commerce conversations across web, SMS, and email. This talk shares the engineering lessons behind building it at scale: multi-tenant RAG pipelines, orchestrator simplification, circuit-breaker gateways, and why manual error analysis beats aggregate metrics every single time.

Elliot Penson
Klaviyo
Watch →
Solutions Track

Beyond Sandboxes: Architecting Durable Runtimes for AI Agents

Sandboxes protect. Durable runtimes survive. This talk argues that production AI agents need more than isolation — they need orchestration-layer infrastructure that persists state, checkpoints every step, and recovers from crashes without losing context. Learn how treating agents as late-bound sagas unlocks retries, human-in-the-loop pausing, multi-agent coordination, and auditability at scale.

Viren Baraiya
Orkes
Watch →
Solutions Track

Building the Context Engine AI Agents Need

AI coding agents aren't failing because they lack intelligence — they're failing because they lack context. This talk introduces the context engine: a reasoning layer that continuously synthesizes organizational knowledge, resolves conflicting sources, and delivers identity-scoped, token-optimized context so your agents produce correct, org-aware code instead of triggering costly correction loops.

Brandon Waselnuk
Unblocked
Watch →
Solutions Track

Making AI Reliable in the SDLC: The Role of Deterministic Context at Scale

LLMs hallucinate architecture. Static analysis doesn't. This talk shows how grounding AI coding agents in deterministic, graph-based architectural context — rather than probabilistic inference — cuts token consumption by 50%, doubles output accuracy, and compressed a seven-year mainframe modernization timeline to just three years. Reliability at enterprise scale isn't magic; it's determinism.

Julien Godfroid & Max Kozinenko
CAST Software · SoftServe
Watch →
Solutions Track

The MCP Megalith: Surviving the 1,000-Tool Explosion and Auth Fragmentation

MCP grew from 100 servers to 30,000 in months. Now comes the hangover: tool sprawl bloating context windows, OAuth fragmented across every vendor, spec reversals breaking pipelines quarterly, and providers renaming tools without notice. This talk is the operational survival guide enterprises need before going all-in on MCP at scale.

Neil Mansilla
Barndoor AI
Watch →
Solutions Track

Make Quality Executable: Verification Gates for Agentic Engineering at Scale

AI writes code faster than humans can review it. This talk argues that verification gates — embedded at planning, commit, CI, and PR stages — are the only way to scale agentic engineering without sacrificing quality. Learn how codified rules, multi-agent review, and continuous feedback loops finally make standards executable, not aspirational.

Nnenna Ndukwe
Qodo
Watch →

About QCon AI Boston 2026

QCon AI is the practitioner-focused conference for engineers building production AI systems. Two days, three tracks, and a single mandate: real lessons from teams shipping agents, retrieval pipelines, and LLM-powered products at scale. InfoQ partners with QCon to bring these recordings to the wider engineering community.

Visit QCon AI Boston More on InfoQ
BT