| |
|
Agents & APIs Demo Night
|
| With Pooja Mistry (Dev Advocate, Postman), Muhammad Annas Hashmi (DevRel, Daytona), Ryan Cooke (Energy Enggr, EastPoint), Abir Taheer (Composio). |
|
Jun 02 (Tue) @ 05:30 PM
FREE
| | Venue, 56 Greene St, 4th Fl
|
|
|
|
|
| |
| Register |
|
|
| |
Kick off Tech Week NY the right way!
Come hang with Composio, Daytona, Postman, Render, WorkOS, & the Tech Week NY developer community for a fun night of drinks, pizza, & tech demos from those building for the future, using Agents & APIs.
This event is a part of #NYTechWeek-a week of events hosted by VCs & startups to bring together the tech ecosystem. Learn more at www.tech-week.com.
Legal Stuff
By registering you agree that listed host companies can send you updates & news about thier products & services. You can unsubscribe anytime using the link in an email or by contacting the companies.
Demo Topics & Speakers
Postman: Pooja Mistry, Sr. Developer Advocate
The Best Time to Build is Now: How an API Tool Became an AI Platform
From collections to MCP servers, agent mode to AI workflows . See how one platform evolved with the ecosystem & why the best time to build is now.
WorkOS: Ryan Cooke, Engineering
Ryan Cooke will give a preview of Studio: a new product from WorkOS, that generates live, interactive miniature apps from natural language. It answers questions with graphs & charts instead of text, reimagining business intelligence as live UIs instead of static dashboards.
Daytona: Muhammad Annas Hashmi, DevRel at Daytona
All tests pass" means nothing, especially if your agent says so.
CI was already a weak signal, but it's even more useless when the same agent wrote both the implementation & the test that's supposed to catch its mistakes.
This demo shows the benefits of making your agents give every PR a receipt. A coding agent boots the PR branch in a Daytona sandbox, drives the running app through a real browser (or via computer-use), & attaches the recording back to GitHub - all of which happens on the cloud.
The same loop applies to AI code review. When a code review agent (Greptile, CodeRabbit, Copilot, or your custom agent) flags a potential bug, the agent tries to reproduce it inside a live sandbox: confirming or refuting the bug with video proof.
Stop being your agent's QA. The work & the proof should arrive together.
Render: Hazal Mestci, Developer Relations Engineer
Parallel AI image workflows, because one model is never enough to disappoint me
This talk demonstrates a Render Workflows-powered system that runs blog thumbnail generation across multiple AI models (OpenAI, Gemini, & DALLE) in isolated compute for parallel inference. It orchestrates execution with retry-safe workflows to coordinate tasks, aggregate results, & handle failures, stores outputs in S3-compatible MinIO, & serves them to a React frontend for side-by-side comparison.
Attendees will learn how to build parallel AI workflows using orchestration, implement reliable fan-out/fan-in patterns for multi-model inference, & connect distributed compute, storage, & APIs into a production-ready end-to-end application.
Abir Taheer, Member of Technical Staff, Composio
Runtime Permissions for Agentic Tools
AI agents are getting access to more tools - but current permission models stop at "allow" or "deny." What happens when a single tool call can trigger actions across multiple services? We'll explore building a permission layer that gives users visibility & control over each action an agent takes on their behalf
|
|
|
|
|
|
|
|
