- unwind ai
- Posts
- Autonomous ML Engineering Agent
Autonomous ML Engineering Agent
PLUS: Study Mode in ChatGPT, Multi-agent AI workforce on your desktop
Shubham Saboo & Gargi Gupta
July 31, 2025
In partnership with
Today’s top AI Highlights:
The first multi-agent workforce on your desktop
Opensource drag-and-drop builder to create LlamaIndex agents
Opensource MoE video model from China
The first agentic Machine Learning engineer
A glamorous AI coding agent for your terminal
& so much more!
Read time: 3 mins
AI Tutorial
Integrating travel services as a developer often means wrestling with a patchwork of inconsistent APIs. Each API - whether for maps, weather, bookings, or calendars - brings its own implementations, auth, and maintenance burdens. The travel industry's fragmented tech landscape creates unnecessary complexity that distracts from building great user experiences.
In this tutorial, we’ll build a multi-agent AI travel planner using MCP servers as universal connectors. By using MCP as a standardized layer, we can focus on creating intelligent agent behaviors rather than API-specific quirks. Our application will orchestrate specialized AI agents that handle different aspects of travel planning while using external services via MCP.
We'll use the Agno framework to create a team of specialized AI agents that collaborate to create comprehensive travel plans, with each agent handling a specific aspect of travel planning - maps, weather, accommodations, and calendar events.
We share hands-on tutorials like this every week, designed to help you stay ahead in the world of AI. If you're serious about leveling up your AI skills and staying ahead of the curve, subscribe now and be the first to access our latest tutorials.
Latest Developments
A complete AI workforce just became available for your desktop, and it operates like hiring a team of specialists who know exactly how to work together.
Eigent is a complete workforce of AI agents, each with a specific skill set, that work in parallel, automatically dividing complex tasks among themselves to achieve a common goal, while you focus on the bigger picture.
Powered by CAMEL AI's multi-agent framework, this desktop application deploys specialized agents that integrate deeply with your existing tools via MCP, connecting seamlessly to Google Suite, Slack, Notion, and even your internal APIs. Build your own specialized worker nodes with custom tools or start directly with these pre-built agents - Developer for coding, Search Agent for web research, Document Agent for content creation, and Multi-modal Agent.
Available now in public beta for macOS and Windows with full opensource access on GitHub.
Key Highlights:
Multi-Agent System - Agents automatically break down complex tasks and assign work to the most suitable team members, handling coordination and handoffs without manual intervention.
Tool Ecosystem - Built-in MCP integrations for web browsing, code execution, and popular productivity tools, plus support for custom tool installation and internal API connections.
Adaptive Problem Solving - When tasks encounter obstacles or uncertainty, the system dynamically adjusts its approach and requests human guidance only when necessary for optimal outcomes.
Deployment Options - Choose between local deployment with your own models for maximum privacy and control, or cloud hosting for instant setup with managed infrastructure. You can use Eigent with local models too.
Choose the Right AI Tools
With thousands of AI tools available, how do you know which ones are worth your money? Subscribe to Mindstream and get our expert guide comparing 40+ popular AI tools. Discover which free options rival paid versions and when upgrading is essential. Stop overspending on tools you don't need and find the perfect AI stack for your workflow.
LlamaIndex just dropped FlowMaker, and it's basically Figma for AI agents.
This opensource drag-and-drop builder lets you create complex agent workflows without touching a single line of code, then exports everything into production-ready TypeScript.
You can connect user inputs to language models, add conditional logic, RAG, and watch your agent come to life right in the browser. The whole thing runs on a dual-execution setup where your visual creation works both as an interactive prototype and a standalone Node.js script. Powered by LlamaIndex under the hood, it's designed for developers who want to prototype fast but still need full control over their agent logic.
Key Highlights:
Visual-to-Code Pipeline - Design workflows with drag-and-drop components, then export matching TypeScript code that runs independently as Node.js scripts.
Live Testing Environment - Run and debug agent workflows directly in the browser with step-by-step execution and optional debug mode for deeper insights.
Smart Agent Logic - Build conditional branching and loop-back mechanisms that let agents route conversations intelligently based on context and user inputs.
Opensource - MIT-licensed platform built on Next.js and TypeScript, ready for customization and feature additions to match your specific needs.
Quick Bites
China continues to make leaps in opensource AI as Alibaba drops Wan2.2, bringing MoE architecture to video generation for the first time in an open-source model. The release includes text-to-video, image-to-video, and unified TI2V models, all generating 720p@24fps videos. You can control lighting, color tone, composition, and camera motion directly through prompts. Running on a single 4090, the models are available on Wan’s website, Hugging Face, and Modelscope.
OpenAI just launched Study Mode for ChatGPT, designed to stop students from getting instant homework solutions. It guides students through problems step-by-step using Socratic questioning and personalized scaffolding, rather than spitting out direct answers. The feature launches today for all user tiers.
The first agentic Machine Learning engineer that medals on Kaggle more than most humans. Meet Neo AI, an autonomous multi-agent system that handles everything from messy data ingestion to production deployment, powered by 11 specialized agents working in parallel.
Chat directly with NEO to define problems, guide experiments, and request specific analyses or model adjustments
Runs 100s of fine-tuning experiments automatically, testing different hyperparameters, architectures, and preprocessing strategies
Real-time tracking of model performance, latency, accuracy, infrastructure health, load balancing, and cost optimization
Scores a 34.2% medal rate on 75 Kaggle competitions, setting a new bar for ML agents.
Tools of the Trade
Crush: Charm's AI coding agent that brings their signature beautiful terminal UI design to the CLI agents with rich text rendering and smooth TUI interactions. Comes with the usual feature set like multi-LLM support, maintaining multiple work sessions and contexts, and MCP integration.
Code Context: A semantic code search engine that indexes entire codebases using vector embeddings for deep codebase context, instead of keyword-based search. Integrates with IDEs, CLI agents, and other tools via MCP.
Zread.ai: Converts a GitHub repository URL into an organized, readable book format. It creates a clear project structure for navigation and generates documentation to explain the system's components. Also includes a "Buzz" feature to surface relevant community insights, commits, and news.
Sitefetch: CLI tool that crawls entire websites and extracts their text content into a single file for AI model consumption. Includes configurable crawling patterns, content selectors for custom extraction, and concurrent fetching for performance.
Awesome LLM Apps: Build awesome LLM apps with RAG, AI agents, MCP, and more to interact with data sources like GitHub, Gmail, PDFs, and YouTube videos, and automate complex work.
Hot Takes
The path to solving hunger, disease and poverty is AI and robotics ~
Elon Musklol imagine running an ai lab and not paying your top researchers $100m smh ngmi ~
Bojan Tunguz
That’s all for today! See you tomorrow with more such AI-filled content.
Don’t forget to share this newsletter on your social channels and tag Unwind AI to support us!
PS: We curate this AI newsletter every day for FREE, your support is what keeps us going. If you find value in what you read, share it with at least one, two (or 20) of your friends 😉
Reply