Composabl

Composabl

Technology, Information and Internet

San Francisco, CA 1,388 followers

Build Intelligent Autonomous Agents

About us

Composabl is a platform for creating intelligent autonomous agents from composable building blocks.

Industry
Technology, Information and Internet
Company size
2-10 employees
Headquarters
San Francisco, CA
Type
Privately Held
Founded
2023

Locations

Employees at Composabl

Updates

  • View organization page for Composabl, graphic

    1,388 followers

    Hello Everyone, Please join us for a LIVE demo! "Climb under the hood" and learn how to build, train, and publish an #AI #agent with the Composabl Platform! Register here ➡️ https://lnkd.in/gaFjjNAA Experience what it's like to build an AI agent that makes million-dollar decisions 13% better than the best industry benchmarks. Walk away knowing how to: -Design multiple agents for a real industrial use case -Use your own Python algorithms and machine learning models in Composabl  -Train agents to perform the high-value tasks you need -Analyze agent behavior -Deploy trained agents to control real equipment PS: If you registered for the Nov 14th session great, do not be confused! That session is full hence this new session ;) #MachineTeaching #MachineLearning #AutonomousAI #Industry40 #Automation #SmartManufacturing #AIforIndustry  

  • View organization page for Composabl, graphic

    1,388 followers

    This Thursday LIVE! Experience what it's like to build an AI agent that makes million-dollar decisions 13% better than the best industry benchmarks. Register here ➡️ https://lnkd.in/gaFjjNAA Walk away knowing how to: -Design multiple agents for a real industrial use case -Use your own Python algorithms and machine learning models in Composabl -Train agents to perform the high-value tasks you need -Analyze agent behavior -Deploy trained agents to control real equipment #MachineTeaching #MachineLearning #AutonomousAI #Industry40 #Automation #SmartManufacturing #AIforIndustry

    • No alternative text description for this image
  • Composabl reposted this

    View profile for Kence Anderson, graphic

    Industrial-Strength Intelligent Autonomous Agents

    Agent orchestration is the glue that holds #ai agents together. It is becoming clear that #intelligentautonomousagents require more than just one technology to perform complex tasks well, so orchestration is very important. The more complex and high-value the task, the more likely that several methods will be needed for your agent to achieve it's goals. Here are just a few of the possible building blocks for intelligent autonomous agents: 🧱 Rules: We're already seeing most agent platforms, even ones that are primarily based on large language models (LLM), add rules, because they are such an important way to transfer expertise about how to perform tasks. 🧱 Learning Algorithms: Some sub-tasks and skills are much easier learned by practicing than outlined by dozens or hundreds of rules and exceptions to rules. And sometimes language is too ambiguous to define and teach complex tasks. 🧱 Machine Learning Models: In the same way that the rods and cones in our eyes provide perception beyond simple light sensors, #ML models can provide perception to your agents (visions, hearing, prediction, detection, etc). 🧱 Physics Based Modeling and Control: modeling and control based on physics and chemistry has been used in production systems for over 100 years. It is especially useful for low-level control of phyical systems like robots. 🧱 Optimization: Often searching options is the best bet for finding the next best action to take. It is also particularly useful for planning. 🧱 Generative AI: I don't have to tell anyone that language models hold great promise for agents, but just like the language areas in our brain (Broca's and Wernicke's area), they are just part of the complete picture of intelligent decision-making. #engineeredintelligence #usefulAI ✅ For more discussion about #ai agents, consider 👍, ✍️, ♻️ and following me on Linkedin: https://lnkd.in/gEXevTCx, then activate the 🔔

    • No alternative text description for this image
  • Composabl reposted this

    View profile for Kence Anderson, graphic

    Industrial-Strength Intelligent Autonomous Agents

    There are three known ways to orchestrate agents to perform tasks: rules, language, and learning. 1️⃣ Agent Orchestration with 📏 Rules📏 We are already starting to see language-based agent libraries and frameworks begin to acknowledge that rules based orchestration is required to perform most enterprise-grade tasks reliably. Folks are inventing all kinds of names and terms to describe it, but we're essentially talking about programmed rules that define the orchestration. ✅ Pros: Enterprise systems have used rules for orchestration for decades. ⛔ Cons: Rules orchestration usually requires writing a lot of code. 2️⃣ Agent Orchestration with 💬 Language 💬 Humans naturally want to communicate with language, so the idea that you can orchestrate agents with language is very attractive. ✅ Pros: Flexible, easy to refine, natural interface for human experts. ⛔ Cons: Language is ambiguous and difficult to drive agents toward goals. 3️⃣ Agent Orchestration with 🧠 Learning 🧠 Using technologies like #reinforcement learning, it is possible to learn when and how to activate (orchestrate) agents together to perform complex tasks. ✅ Pros: Most flexible way to orchestrate agents to achieve goals (optimize). ⛔ Cons: Learning can take a long time and requires up-front design. The Composabl platform supports all three forms of agent orchestration, so you don't have to choose one over the others, but it is important to understand the differences as you design your agentic systems! #intelligentagents #engineeredintelligence #usefulAI ✅ For more discussion about #ai agents, consider 👍, ✍️, ♻️ and following me on Linkedin: https://lnkd.in/gEXevTCx, then activate the 🔔

    • No alternative text description for this image
  • Composabl reposted this

    View profile for Kence Anderson, graphic

    Industrial-Strength Intelligent Autonomous Agents

    Pretending that a language model (#llm) can make high-value decisions because it read the internet is like me pretending that I know how to work in a steel mill. Expertise matters. I built Composabl to give engineers the choice and power to orchestrate the technologies into agents that best control their high-value equipment and processes. They designed the equipment in the first place, let them decide which algorithms and #ai models to use in the agents that control them. But large language models provide an exceptional natural language interface for decision-making agents. Take a look at these agent components: 📢 The Analyst is an agent pattern template that uses an LLM to explain agent behavior. A plant operator might ask "what is the agent doing" and the LLM might respond with charts and text that explains that the agent is acting in agreement with standard operating procedures. ➡️ The Plant Manager is an agent pattern template that uses an LLM to translate human commands into variables that influence decision-making algorithms. A human operator might say "this equipment is running too hot" and the Plant Manager LLM would output setpoints that alter the temperature of the reactor. 🧠 The Executive is an agent pattern template that uses an LLM to research information that helps the agent make better decisions. For example, the Executive LLM might research market prices of input materials and recommend to increase production. The result is a matrix of algorithms, agents, and co-pilot assistants that work together with humans to make high-value decisions at expert level. #intelligentagents #autonomousAI #usefulAI ✅ For more discussion about #ai agents, consider 👍, ✍️, ♻️ and following me on Linkedin: https://lnkd.in/gEXevTCx, then activate the 🔔

    • No alternative text description for this image
  • Composabl reposted this

    View profile for Kence Anderson, graphic

    Industrial-Strength Intelligent Autonomous Agents

    How many files does your agent have to move around or chat support requests does it have to answer to generate $1M USD in ROI? Every intelligent autonomous agent I've ever designed for manufacturing and logistics generates at least $1M USD in ROI to make a single decision 1% - 10% better. Higher value #ai agent applications require high reliability for extremely nuanced and complex decisions. These complex, high-value decisions cannot be made by jamming in a single algorithm into the process. In fact, only by combining enabling technologies (reinforcementlearning, ML, generativeAI, knowledgegraphs) can you make these decisions at expert human level. What is your agent doing, exactly? Is your agent an expert? #intelligentautonomousagents #engineeredintelligence #machineteaching

    • No alternative text description for this image
  • View organization page for Composabl, graphic

    1,388 followers

    You like to build, tinker and solve million-dollar problems right? Take a look under the hood at this #ProductionScheduling use case where #engineers can experience how #MachineTeaching truly comes to life. >9 of 13 Composabl agents beat the benchmark >Final Result: An agent that delivered a 21% increase profit margin over the current ROI benchmark of 7%. Say whaaaaaaaaatttt?! Read this and other use cases here: https://lnkd.in/gcFWEGnr

Similar pages

Funding

Composabl 1 total round

Last Round

Seed

US$ 5.7M

See more info on crunchbase