Anjaneya Sastry Kappagantu is a Solution Architect at AT&T working on local AI environments and MLX. At Agent Conf, he’ll explore three ways models can adapt during inference, and how to decide when feedback should trigger an update. Tickets: www.agent.sh/tickets

A green diff can still ship a broken mobile app. When agents write more of the code, review needs a third layer: behavior proof from the running app. @thymikee.bsky.social breaks down how you can build your own QA agent with device control & cloud execution fit into that workflow ➡️ clstk.com/4h8MU1J

We’re proud to welcome Amazon as the Platinum Sponsor of Agent Conf 2026. The Amazon team will also join the program, with Anisha Malde and Kourtney Meiss speaking about harness engineering and multi-agent orchestration. Excited to have you on board! 👉 www.agent.sh

Assess the product before choosing a migration path. Review enough code, delivery, and live app behavior to catch a poor fit before it becomes cutover risk, rework, or years of maintaining two runtimes. Leave with trial flows and go/no-go criteria: 👉 clstk.com/4wHuR7M

Building camera features in React Native? This episode of React Universe On Air with @mrousavy.com covers what changed in VisionCamera v5: Nitro Modules, the constraints API, output objects, migration from v4, and the testing work behind the release. Get show notes and resources at clstk.com/45BOQZn

Carl Ross is a Member of Technical Staff at OpenAI, working with agentic software factories across the software development lifecycle. Now he’s joining Agent Conf, where he’ll cover what happens when token budget is entirely removed as a constraint. Get your ticket: agent.sh/tickets

Susmita Singh is a Senior Software Engineer at Google who led AI infrastructure work for the Google Store. Now she's joining Agent Conf in Warsaw as a Speaker. In her talk she’ll cover how to build grounded and reliable AI agents at scale. Tickets: www.agent.sh/tickets

Agent Conf 2026 | Warsaw | Shaping the Future of Agentic Development

Join 500+ engineers in Warsaw Sept 17-18 2026 for Agent Conf. Learn agentic development, orchestration, and AI-driven workflows. Hosted by Callstack.

agent.sh

Brownfield keeps the current app as the production host. Greenfield builds the replacement alongside it. A checkpoint path lets teams test greenfield before committing. The choice depends on where the migration risk sits in the product. Read the article: clstk.com/4fAxKkI

An agent took 3:18 to complete a 9-step Android QA flow. The same recorded route replayed in 8.8s, including launch, with zero model calls. We ran it 20 times to be sure. Are agents wasting that much time? @thymikee.bsky.social explains the tradeoffs ➡️ clstk.com/4c6MkOr

A React Native migration changes how teams build, release, and measure mobile work. Our new ebook covers brownfield, greenfield, a greenfield-first option with a tested fallback, the two-pass AI workflow, release control, and ROI. Download the guide: clstk.com/4pKVK7Y

Five years of VisionCamera work led to this release. In the upcoming podcast, Mike talks with Marc Rousavy about what changed under the hood, how Nitro shaped v5, and why this release goes far beyond a routine upgrade. Premiering on August 4 ➡️ youtu.be/xI3egb3_ihI

React Native updates arrive through releases, RFCs, repositories, and the wider ecosystem. Join Kewin and Hur live as they build a news agent with Eve: configure a custom model, send updates to Slack, deploy to Vercel, and inspect its activity ➡️ clstk.com/4yHt9o3

Bundle size is one of the topics teams keep postponing until users are hit with “large size” warning when downloading your app and stop the process. Our guide breaks down what is worth checking: dependency cost, barrel exports, tree shaking, platform shaking & bundling choices that affect startup.

AI can produce a plausible PR quickly, but the merge decision still depends on how the changed app behaves on a real device. Our webinar covers behavior review: bounded QA agents, runtime evidence, and a practical first workflow. Link to recording ➡️ clstk.com/4vDO6xv

What does AI look like when it leaves chat window? Lech built Vault, a handheld terminal for self-hosted agents and local models. Its OLED tracks live task activity & e-paper holds the finished answer. Read full story ➡️ clstk.com/4vLuKXt OpenClaw, DGX Spark, TensorRT-LLM. Two screens, one task.

Our new guide builds the smallest working React Native super-app setup with Re.Pack 5 and Module Federation. You will configure a host and mini app, expose a remote module, load it with React.lazy, and add loading and error states around the remote screen ➡️ clstk.com/3RQj9IF

Give Metrognome a slow React Native screen and a night. It proposes a fix, drives the app, profiles it, and keeps the changes that hold up. Xavier Costa from Uphold explains how it uses Agent Device, Agent React DevTools, Metro MCP, and React Native Best Practices ➡️ clstk.com/4gIgM4M

Simple motion should not always require complex machinery. Łukasz Chludziński & Janic Duplessis talk about React Native Ease, simple transitions, Core Animation on iOS, where Reanimated still wins, and why less work can matter for resource use and battery life. Watch full episode: clstk.com/4hb8J0w

For one of our clients, we built a design-to-code workflow with enforced component usage and a visual QA loop, so generated screens had to match the source design before handoff. That helped the team prototype faster and control UI quality across web and mobile. Read case study: clstk.com/4fm1Net

A plausible PR is not the same thing as a safe release. If your review still stops at the diff, this webinar will show what is missing before the merge. Register to join us live on July 21 and see what that review loop can include: clstk.com/4vXe0xb

Full house in Dubai. 🥳 Today we’re meeting for the OpenAI Codex Community Meetup, and we’re excited to spend the evening with a room full of builders working with Codex and coding agents. More about the Meetup 👉 clstk.com/4ppFwkM

Support AI adoption in your company by sending your engineers to Agent Conf. Two days on prompting, debugging agent behavior, multi-agent workflows, and adoption inside real teams. Use the promo codeTeams15 to get 15% off 3+ tickets. Get tickets: agent.sh/tickets

Native React Native debugging gets much easier once you ask where the work runs. Sync Turbo Module calls can hit the JS thread. Async calls move to the Turbo Modules thread. Fabric view updates stay on the main thread. That context saves time.

David Budáč has worked with Oracle Database for 20 years as a DBA, developer, and consultant. Now he’s joining Agent Conf in Warsaw as a speaker. He’ll talk about using agents to analyze logs and build a queryable memory of operational events. Get your ticket: www.agent.sh/tickets

iOS packaging time was burning close to half an houron every Brownfield build. We cut that by up to 39% by using React Native prebuilts in the packaging flow in `@callstack/react-native-brownfield` 3.10. Less waiting in CI. Faster iteration on hybrid apps. Read the article 👇

This time, we talk with @lahteenlahti.com about how AI is starting to change monetization tooling. We get into AI-generated paywalls, MCP, experiment suggestions inside the dashboard, and what is coming next for onboarding flows, ads, virtual currencies, and React Native Web ➡️ clstk.com/4arkcFg

What AI Changes for Paywalls and Monetization

Perttu Lähteenlahti (RevenueCat) talks about AI-generated paywalls, MCP tooling, monetization experiments, and what RevenueCat is building next across mobile and web.

clstk.com

Some React Native animations can take a lighter path. New React Universe On Air is live: Łukasz talks with Janic from App & Flow about React Native Ease, Core Animation on iOS, where Reanimated still wins, and why doing less work can be a real performance advantage. Watch now

Review capacity isn't scaling with PR volume for your team? Join our webinar, where we will show what to add before merge when code review and E2E are not enough. Register below for the event on Jul 21 ➡️ clstk.com/3SPRUyb

Nishant Gupta is a Staff Software Engineer and Researcher at Meta, focused on large-scale distributed systems and applied AI. Now he’s joining Agent Conf in Warsaw as a speaker. He’ll talk about managing thousands of AI agents without losing control. Get your ticket: www.agent.sh/tickets

At @bookingcom.bsky.social, Huang Minghe works on GraphQL Federation and the Agent Skills ecosystem. Now he’s joining Agent Conf in Warsaw as a speaker. He’ll talk about autonomous migration inside a 40-million-line legacy codebase. Get your ticket 👉 www.agent.sh/tickets