←
Chef Robotics340 posts
Chef Robotics avatar

Chef Robotics

@ChefRobotics

Empowering humans to do what humans do best. We're hiring: chefrobotics.ai/careers

◷ Joined May 2018▧ 170 Photos and videos
Chef Robotics avatar
Chef Robotics@ChefRobotics·20h

Chef robots specialize in placing food into a tray, but food manufacturers still need to get each tray sealed, wrapped, and ready to ship. We're excited to announce our partnership with @ameri_pak to help food manufacturers automate both meal assembly and packaging together. 𝗙𝗿𝗼𝗺 𝗶𝗻𝗴𝗿𝗲𝗱𝗶𝗲𝗻𝘁 𝗽𝗹𝗮𝗰𝗲𝗺𝗲𝗻𝘁 𝘁𝗼 𝗮 𝘀𝗲𝗮𝗹𝗲𝗱 𝗽𝗮𝗰𝗸𝗮𝗴𝗲 Chef's AI-enabled robots pick, portion, and place food ingredients into meal trays and onto wraps, adapting to different ingredients, portion sizes, and tray formats. AmeriPak provides tray sealers, flow wrappers, and shrink wrappers to wrap those trays. Together, we'll help manufacturers make a smooth transition between these two production stages, no matter the tray format, line layouts, or throughput goal. 𝗕𝘂𝗶𝗹𝘁 𝗮𝗿𝗼𝘂𝗻𝗱 𝗰𝗵𝗮𝗻𝗴𝗶𝗻𝗴 𝗺𝗲𝗻𝘂𝘀 Chef robots and AmeriPak's packaging equipment are built for manufacturers that run many SKUs and change their menus frequently, enabling them to increase output without losing flexibility. Manufacturers can start the conversation with either team and connect with the other, while continuing to work directly with each company for equipment and support. Learn more about our partnership with AmeriPak and what it means for food manufacturers: chefrobotics.ai/post/chef-ro…

0:50
Chef Robotics avatar
Chef Robotics@ChefRobotics·Nov 16

A robot doesn’t need to do an entire human job to be useful. It needs to do a specific task exceptionally well. Our CTO and founding engineer, Luis Rayas, joined the FuturePulse podcast to explain how this shapes our work at Chef. For food assembly, that means fast, precise, consistent portioning, freeing people to solve problems and handle unexpected situations on the production line. This also shapes how we think about general-purpose robotics. An excavator’s shape makes sense for digging; giving it a human body wouldn’t make it a better tool. Luis sees more promise in a general-purpose robot brain that learns to manipulate different things, with hardware suited to each application. At Chef, deploying robots teaches us which problems actually need learning and where a simpler approach works better. Our growing fleet also generates data that makes new learning approaches possible. That’s what “deployment first” means at Chef: customer needs shape the technology, and production experience shapes what we build next. Thanks, Melica Moshiri, for having Luis! Watch the full conversation: youtube.com/watch?v=OhEATMcD… #PhysicalAI #foodautomation #AIandrobotics

Chef Robotics avatar
Chef Robotics@ChefRobotics·Nov 5

Working with a large customer can give a robotics company the funding, production experience, and orders it needs to grow. But what happens when that customer slows down or finishes its rollout? The Robot Report explores this question through Symbotic’s relationship with Walmart. One thing that stood out to us was the recommendation to involve several customers early in product development so their requirements help shape your system from the beginning. In food manufacturing, for example, manufacturers may work with different ingredients, portion sizes, and production line layouts. Understanding those differences early helps a team decide what to build into the core system and what to make configurable. Early deployments also show how much engineering and on-site work each new customer requires. Knowing that helps us plan how to serve new customers while supporting existing ones. Read Deepak Jayaraj's article to learn more about managing customer dependence and planning for growth: therobotreport.com/what-symb…

Post attachment
Chef Robotics avatar
Chef Robotics@ChefRobotics·Nov 3

Sam Padilla and Varun Nair's robotics map gave us a new perspective by depicting the robotics industry as a subway system. Subway stations represent the industry's layers (e.g., robots, models, deployments, data services). Companies that span more than one layer are placed on one or more subway lines between the stations. The busiest connection, not unexpectedly, is the deployments ↔ robots line. Some companies, like Dyna Robotics, ride more than one subway line, and NVIDIA is represented at four different subway stations. Beyond the visual representation of how different layers of the robotics industry are connected and which companies are at the forefront of each, we're interested to see how these layers and connections will change over time and how the pendulum between robots and deployments, models and simulation, data tools and data services will swing over time. Explore the map and let us know what you find interesting about it: roboticsmap.org/c/chef-robot…

Post attachment
Chef Robotics avatar
Chef Robotics@ChefRobotics·Nov 2

Part 1 of our Food World Model blog series explored how our model predicts the weight of a candidate pick and the surface it will leave behind in an ingredient pan. Today, we’re taking a look at where our training data comes from and how we turn production robot data into a scalable training flywheel. Getting the training flywheel right is crucial because with world models, it's not enough to predict a plausible future. A world model's predictions need to hold up against physical measurements and, in our case, across a range of food ingredients and production conditions. Our AI research engineering team turned 150 million production robot actions into a reproducible data corpus more than 10x larger than the one mentioned in Part 1, and each new training run is cheaper than the previous one. Read more: chefrobotics.ai/post/buildin…

Post attachment
Chef Robotics avatar
Chef Robotics@ChefRobotics·Oct 29

Taking the PACK EXPO conversations out on the water! 🛥️ We hosted a boat cruise in Chicago during PACK EXPO to bring food manufacturing folks together off the busy show floor and spend more time getting to know one another. Trade shows bring so many people into the same city, but there’s rarely enough time for every conversation. We’re glad we could make some room for that. Thank you to everyone who joined us and made the cruise part of their time in Chicago. It was a pleasure having you aboard! If we didn’t get to connect in Chicago, we’d love to continue the conversation: chefrobotics.ai/contact-us?u… #PACKEXPO #foodautomation #physicalAI #foodmanufacturing

Post attachmentPost attachment
Jeson Lee avatar
⟳ Chef Robotics retweeted
Jeson Lee@thejesonlee·Oct 23

Last night, @savant_vc and @scalevp brought together 20 reporters, investors and founders at Birdsong to talk about where Physical AI is headed. We spent a lot of the night on what this wave can learn from self-driving and LLMs, how the market might evolve, and what we’re seeing on the ground. Always great to bring together Savant portfolio founders & Residents: @PrannayKhosla (Guild Robotics), @aurorafeng_01 (Neural Motion), Da Fang (Shepherd Robotics), Gabe Rodriguez (Puppet Robotics), Jimmy Jin, Arun Venkatraman & Nikhil J Joshi (stealth) Alongside investors: Siddharth Ramakrishnan (@scalevp), Vikram Ramakrishnan (A*), @CorinneMRiley (@GreylockVC), @McCreadieWill (@redglassvc), @Chengdavid923 (@costanoavc) And reporters and leaders in Physical AI: @ryajetha (@BusinessInsider), @geneteare (@crunchbasenews), @RachylJones (@sfstandard), @E0M (@Generalistai), Luis Rayas (@ChefRobotics) Thanks so much @nehajewalikar for co-hosting this with us and making it an A+ experience for everyone. If you’re interested in a dinner like this, reply “dine me” here to get on the invite list!

Post attachment
Chef Robotics avatar
Chef Robotics@ChefRobotics·Oct 18

We're at PACK EXPO International in Chicago from today through Oct 21! Stop by 𝗯𝗼𝗼𝘁𝗵 𝗟𝗟-𝟭𝟬𝟰𝟬𝟭 to see how our robot handles food assembly, from picking ingredients to portioning them into trays. We’re also hosting a happy hour tomorrow at our booth from 4-6 PM CT. Connect with food manufacturers, co-packers, packaging equipment manufacturers, and automation partners over drinks. RSVP here: luma.com/hlydxfa0 #PACKEXPO #FoodManufacturing #FoodAutomation #PackagingAutomation

Post attachment
Chef Robotics avatar
Chef Robotics@ChefRobotics·Oct 15

Meet the Chef team at PACK EXPO International in Chicago! Nicolas De Keijser, Joe Weber, Nick Y., Jeff O'Neil, Bennett Dixon, Atharva Sonanis, and Mohamed Elzaki will be there to discuss how Chef's AI-enabled robots can fit into your production line. Stop by to see our robot in action and talk through your production challenges. - 𝗗𝗮𝘁𝗲: 𝗢𝗰𝘁𝗼𝗯𝗲𝗿 𝟭𝟴-𝟮𝟭 - 𝗕𝗼𝗼𝘁𝗵: 𝗟𝗟-𝟭𝟬𝟰𝟬𝟭 We’re also bringing together food manufacturers, co-packers, packaging equipment manufacturers, and automation partners for a happy hour at our booth on Monday, Oct 19, from 4-6 PM CT. Apply for a spot here: luma.com/hlydxfa0 #PACKEXPO #FoodManufacturing #FoodAutomation #PackagingAutomation

1:53
Ilir Aliu avatar
⟳ Chef Robotics retweeted
Ilir Aliu@IlirAliu_·Oct 14

"If your throughput is 50% slower, you get fired [...] If your reliability is 50%, 60%, 70%, you don't even get a sales call." @RajatBhageria, Founder & CEO of @ChefRobotics, has a different perspective on the current humanoid robotics race. The industry is betting heavily on general-purpose robots. The assumption is that once a robot can perform almost any task, we've solved robotics. But industrial customers have different priorities. In manufacturing, throughput and reliability determine whether a robot makes it onto the factory floor. A robot that performs 100 different tasks at 70% reliability may be less valuable than one that performs a single task at 99.9%. Rajat argues that performance and reliability matter more than task generalization in industrial applications. We discussed this and much more on the 22Astronauts Podcast. This episode is brought to you by @CoreWeave, providing the specialized cloud infrastructure that powers physical AI and large-scale model training. Full episode: YouTube: youtu.be/BkfULtOh5hU Spotify: open.spotify.com/episode/0Ae… 𝐓𝐈𝐌𝐄𝐒𝐓𝐀𝐌𝐏𝐒 (00:00) Why Chef walked away from restaurant deals (03:39) Carnegie, Rockefeller, Edison (08:33) ThirdEye, computer vision for the blind (15:14) Customers don't want cooking robots (18:05) Airline catering, meatpacking, CPG (25:34) The founder's job is persuasion (36:08) Home Depot, not Costco (46:28) Production data vs. simulation (52:52) The bottleneck is the only thing that matters

1:00
Chef Robotics avatar
Chef Robotics@ChefRobotics·Oct 14

We're excited to partner with @Physical_int!

Announcing the beta release of the Pi API. We’ve partnered with companies across diverse industries to adapt our models to their robots and applications, from harvesting strawberries to assembling servers. Now we’re opening access to more teams.

0:35
Chef Robotics avatar
Chef Robotics@ChefRobotics·Oct 13

We’re excited to announce our research partnership with @physical_int to explore food assembly for high-mix, lower-volume operations such as ghost kitchens, fast-casual restaurants, and other commercial kitchens. 𝗪𝗵𝘆 𝗹𝗼𝘄𝗲𝗿-𝘃𝗼𝗹𝘂𝗺𝗲 𝗸𝗶𝘁𝗰𝗵𝗲𝗻𝘀 𝗻𝗲𝗲𝗱 𝗮 𝗱𝗶𝗳𝗳𝗲𝗿𝗲𝗻𝘁 𝗮𝗽𝗽𝗿𝗼𝗮𝗰𝗵 Today, Chef robots help high-mix, high-volume food manufacturers overcome labor shortages. On these lines, each worker handles one ingredient, and throughput and domain flexibility are key. In lower-volume operations such as restaurants, one worker may assemble an entire burger, bowl, or sandwich. Handling every ingredient and step requires more dexterity and flexibility across tasks. 𝗪𝗵𝗮𝘁 𝘄𝗲’𝗿𝗲 𝗿𝗲𝘀𝗲𝗮𝗿𝗰𝗵𝗶𝗻𝗴 𝘄𝗶𝘁𝗵 𝗣𝗜? To unlock these tasks, we’re exploring PI’s pretrained vision-language-action (VLA) models as a base we can post-train for food manipulation. Our bimanual food assembly system learns varied tasks from expert human demonstrations, with two arms working together to assemble a meal. 𝗪𝗵𝗮𝘁 𝘁𝗵𝗶𝘀 𝗰𝗼𝘂𝗹𝗱 𝗲𝗻𝗮𝗯𝗹𝗲? This research could bring automation in food assembly to more settings where flexibility and dexterity matter most, including ghost kitchens, fast casuals, QSRs, hotels, stadiums, and school and university kitchens. Learn more about our partnership: pi.website/blog/api-beta #PhysicalAI #FoodRobotics #Robotics #Physicalintelligence

Post attachment
Chef Robotics avatar
Chef Robotics@ChefRobotics·Oct 12

We’re hiring a Robotic Operations Engineer at Chef! As we scale our deployed robot fleet, we’re looking for someone who enjoys solving technical problems in real time and wants to help bring AI-enabled robots to more food production facilities. You’ll be the first line of support for our robots in the field: troubleshooting remotely, guiding on-site teams, and collaborating with our engineers on escalations. If you have a background in robotics, ROS, or Linux and this sounds like your kind of role, we’d love to hear from you. Learn more and apply here: jobs.gem.com/chef-robotics/a…

Chef Robotics avatar
Chef Robotics@ChefRobotics·Oct 7

We’re excited to share that we are using NVIDIA Robotics software to develop and deploy our robots faster 🤖 𝗧𝗵𝗲 𝗰𝗵𝗮𝗹𝗹𝗲𝗻𝗴𝗲 Food manufacturers work with highly variable ingredients on high-speed lines in cold, harsh conditions. Automating these tasks requires robots that can handle this variability reliably and be deployed quickly in production. 𝗛𝗼𝘄 𝗖𝗵𝗲𝗳 𝘂𝘀𝗲𝘀 𝗡𝗩𝗜𝗗𝗜𝗔 𝘀𝗼𝗳𝘁𝘄𝗮𝗿𝗲 → NVIDIA Isaac Sim, built on NVIDIA Omniverse libraries, allows us to create physically accurate digital twins of our robotic cells and customers’ production lines. Our team can design, simulate, and stress-test configurations before deployment. → NVIDIA cuMotion helps our robots plan fast, collision-free movements using GPU acceleration for individual picks, multiple deposits, and multi-robot operations. 𝗧𝗵𝗲 𝗿𝗲𝘀𝘂𝗹𝘁 These technologies help us shorten development cycles and get new robots into production faster while maintaining reliability. Learn more about how we’re using them: chefrobotics.ai/post/chef-ro… #PhysicalAI #FoodManufacturing #Robotics #FoodAutomation #NVIDIA

Post attachment
Chef Robotics avatar
Chef Robotics@ChefRobotics·Oct 2

How do Chef robots support worker safety on meal assembly lines? Meal assembly involves repetitive scooping, portioning, and spreading, often for hours in 34°F cold production environments. These tasks can contribute to fatigue and repetitive strain injuries. 𝗗𝗲𝘀𝗶𝗴𝗻𝗲𝗱 𝘁𝗼 𝘄𝗼𝗿𝗸 𝗮𝗹𝗼𝗻𝗴𝘀𝗶𝗱𝗲 𝗽𝗲𝗼𝗽𝗹𝗲 Chef robots use collaborative robot arms and are designed using ISO/TS 15066:2016 guidance. Built-in safeguards include protective stops, defined arm-motion limits, and tooling designed to reduce pinch and shear hazards. 𝗥𝗲𝗱𝘂𝗰𝗶𝗻𝗴 𝗽𝗵𝘆𝘀𝗶𝗰𝗮𝗹 𝘀𝘁𝗿𝗮𝗶𝗻 Our robots take on repetitive meal assembly tasks, allowing workers to focus on cooking, quality control, and line management. Ergonomic features also simplify everyday operation: → Slide-in ingredient pans reduce lifting during refills → Self-leveling feet eliminate manual wrench adjustments → Glove-friendly touchscreens simplify operation in cold rooms Read more on how Chef robots support worker safety: chefrobotics.ai/post/how-che… #foodmanufacturing #foodautomation #workersafety #robotics

Post attachment