chai-research-com

Strategic Round II Complete READ MORE 1.4 EXAFLOP CLUSTER AI Characters & Stories Strategic Round II Complete: CHAI AI secured additional funding rounds, bringing its total investment to date to over $55M dollars to expand both compute and talent acquisition. READ MORE ROADMAP [ CHAI Revenue Growth ] Incentives & Scale JOBS We have built a consumer platform where users can build their own AI characters and stories. Our B2C business has grown exponentially, reaching $70M/yr in revenue in 2026. OCT 2022 APR 2023 OCT 2023 APR 2024 OCT 2024 APR 2025 OCT 2025 FEB 2026 NOV 2022 CHAI Launches on App Store We were the first to launch a consumer AI platform, allowing users to create their own ChatAIs—ahead of Character AI and ChatGPT. FEB 2023 Deploys First In-House 6B LLM Open-sourced LLMs no longer satisfied our users' requirements, as the LLMs needed to be adapted for social and engagement purposes. We saw a +10% engagement boost from our own in-house model. MAR 2023 Deploys Best-of-4 Reward Model We continued to iterate on RLHF (Reinforcement Learning with Human Feedback), training a reward model directly on user signals. This led to a huge boost in our day 30 user retention. APR 2023 Larger Model Upgrade - 13B Architecture We found that a bigger model leads to better depth, therefore better retention. We re-trained our LLM from scratch and saw another +10% engagement boost. MAY 2023 PPO Model Deployed Using Proximal Policy Optimization, a reinforcement learning technique, we optimized our base foundation model to decrease the probability a chat session ends. JUNE 2023 Deploys Reward Model XL Continued to scale up our reward model. Trained with 100 million signals to decrease user retry rate and increase chat session length. OCT 2023 Efficient Inference & Custom GPU Orchestration Off-the-shelf load balancing and vLLM were no longer sufficient to support our user base at 500K DAU scale. We implemented custom CUDA kernels together with our own GPU orchestration system. NOV 2023 Increased GPU Reservation We hit a scaling issue due to high demand from our users. We reserved an additional 1,000 A100 GPUs from our provider to scale reliably. NOV 2023 Deployed Model Blending CHAI invented model blending—ensembling different LLMs trained on different targets at the conversation level. This outperformed GPT-3 in user retention. DEC 2023 BO8 Reward Model Deployed With increased cluster capacity, we implemented Best-of-8 rejection sampling, utilizing our upgraded reward model to its full extent. MAR 2024 DPO Model Deployed Utilizing Direct Preference Optimization with user preference datasets, we boosted engagement by 20%. The performance stacked well with our existing reward model. AUG 2024 Upgraded All Existing Blends to DPO Building on the success of DPO, we iterated on optimization targets and data selection, and successfully deployed DPO across all production blends. SEP 2024 13B Reward Model Deployed With increased GPU capacity due to cluster upgrades, we were able to serve larger reward models for all users. OCT 2024 10x 24B Models Deployed We upgraded our existing production blend to 24B models. With blending enabled, we saw a surge in daily active users and day 30 retention. JAN 2025 Model Mesh Orchestrator Deployed To support over 1M Daily Active Users, Model Mesh—an in-house cluster orchestration platform—was deployed to handle multi-cluster, multi-GPU-type serving of hundreds of LLMs in production. MAR 2025 GRPO Deployed GRPO (Group Relative Policy Optimization) is an upgrade from Direct Preference Optimization, resulting in a +15% engagement improvement. SEP 2025 AMD Instinct™ MI Series Deployed All AMD MI325X and MI300X nodes are online and added to Model Mesh, serving over 50% of our production traffic. NOV 2025 Mixture-of-Experts Training Begins We began incorporating large Mixture-of-Experts models into our RLHF research, with the goal of increasing user revenue and screentime through our expanded compute capacity. DEC 2025 2x 235B A22B Deployed We upgrade our production blend to incorporate 235B A22B Mixture-of-Experts models. Platform screentime and both revenue increased by 25%. DEC 2025 2x 235B A22B Deployed We upgrade our production blend to incorporate 235B A22B Mixture-of-Experts models. Platform screentime and both revenue increased by 25%. Growth STAGE 0M+ FUNDING RAISED 0M+ ACTIVE USERS BACKED BY [ GPU Cluster ] 1.4 EXAFLOPS GPU CLUSTERFOR AI INFERENCE At CHAI, we serve hundreds of in-house trained LLMs across several GPU chip types from both AMD and Nvidia. While open-source solutions such as vLLM work well for simple workloads, we've found that we can further improve upon vLLM by almost an order of magnitude through several optimizations, such as custom kernels and compute-efficient attention approximations. NUMBER OF GPUS 5000 GPUs NUMBER OF TOKENS SERVED 1.2T Tokens / Day NUMBER OF UNIQUE LLMS SERVED 51K LLMs CLUSTER COMPUTE PERFORMANCE >1.4 Exaflops NVIDIA A100 NVIDIA L40S AMD Mi325x AMD Mi300x [ Product ] Building Platform for Social AI NEWS We believe in platforms. There is huge demand for AI that is not only factually correct but also entertaining and social. IOS ANDROID CHAI WEB Coming Soon... --- JULY 2, 2025 CHAI SECURES INVESTMENT CHAI AI secured additional rounds of funding, bringing its total investment to date to over $55M. CHAI’s investors include individual investors such as Ken Howery and Bill Elmore, along with AI companies CoreWeave and AMD CHAI's progress is accelerating rapidly. We have secured additional rounds of funding, bringing our total investment to date to over $55M. Our investors include individual investors such as Ken Howery and Bill Elmore, along with AI companies CoreWeave and AMD. With a mission to build a platform for user-generated AI (UGAI), CHAI was the first to enter the UGAI space, launching in 2021 and hitting 1 million users in early 2022, preceding both ChatGPT and Character AI. As of 2025, CHAI has continued its growth and now serves over 10M users and is one of the most valuable private AI startups in Silicon Valley. With its office in Palo Alto and a lean team of 11 engineers, it has an ARR of $40M. "This investment will be spent on doubling the amount of compute per user, empowering us to serve cutting-edge AI to our users." — William Beauchamp, Founder Since our Strategic Round I in January 2024, CHAI has made significant technical progress and rapidly expanded its user base including: Revenue Growth: CHAI has doubled its annualized revenue from $20M to $40M; with improved monetization strategies, we are able to reinvest earnings into our compute infrastructure and talent acquisition. Active User Growth: our Daily Active Users have grown from 800K to 2M while average session time increased from 60 to 90 minutes per active session, surpassing traditional social media platforms such as Youtube and Instagram. Infrastructure Scale: we have quadrupled our GPU cluster capacity to serve over 1.2 trillion tokens per day, enabling us to support our growing user base while enabling researchers to rapidly iterate on new model architectures via fast deployment and AB testing to existing users. CHAI 10x24B Blended Model: our proprietary ensemble approach combining multiple specialized 24B parameter models delivers over 60% user engagement improvement compared to GPT4o. Model Mesh: our custom inference service orchestration platform built to support our growing user base and compute cluster. Enabling us to serve over 10K requests per second across different GPU providers and model architectures. CHAI is expanding rapidly and seeks Engineers and Researchers possessing high levels of craftsmanship ready to join an elite engineering team of the highest talent density that is focused on building the future of Social AI. Those interested can apply today at chai-research.com/jobs. --- [ WHAT'S NEW ] LATEST NEWS Read about our latest product and research announcements JULY 2, 2025 CHAI Completes Strategic Round II CHAI AI secured additional funding rounds, bringing its total investment to date to over $55M dollars to expand both compute and talent acquisition. VIEW STRATEGIC ROUND II PODCAST CHAI x Latent Space Podcast CHAI's founder & CEO William Beauchamp shares his journey from quant trader to building an AI platform. JAN 27, 2025 VIEW RESEARCH Blending is All You Need Our latest research on novel blending techniques for large language models shows 30% improvement in user retention. JAN 03, 2024 VIEW TALK Making a Good LLM - CHAI RLHF Development Cycle Our AI Researcher talks about the development cycle of our RLHF models, optimizing for user retention. NOV 19, 2024 VIEW RESEARCH Reward Modelling for Real-World Engagement with Millions of Users Reward signals are key to the success of RLHF models. We share our learnings from building a model for real-world engagement. AUG 8, 2023 VIEW NEWS Strategic Round II Complete CHAI AI secured additional funding rounds, bringing its total investment to date to over $55M to expand both compute and talent acquisition. JUL 2, 2025 VIEW PODCAST CHAI x Machine Learning Street Talk Our Founder & CEO William Beauchamp talks with Tim from MLST on building an AI startup and hiring a team of hardcore researchers & engineers. JAN 8, 2024 VIEW --- Hiring in Palo Alto Fast growing, High Revenue, Gen AI. Interview process Please send your resume to will@chai-research.com. Overview We run a fast process — an offer can be given within 7 days of first contact. We have no leetcode, behavioural interviews, or design interviews. Assessments are done via take-home and onsite coding challenges, both of which are project-based to reflect an engineer's or researcher's daily work at CHAI. Step 1 Initial chat We do not use recruiters for screening. Instead, an engineer or our founder will have a 20-minute initial video call with you. Step 2 Personality assessment We'll send you a personality assessment to understand you better. This is the one we use. Step 3 Take home challenge We'll send you a short take home challenge that typically takes 60-90 minutes to complete. Step 4 Onsite interview You'll be invited onsite to meet the team and complete the second half of the take home. Step 5 Offer CHAI offers are all-cash. We will offer at least 1.5x over existing TC for all candidates. For example, an SDE II on $300k will get an offer of at least $450k/year salary + cash bonus. Who we are NEWS William BeauchampFounder / CEOCambridge UniversityRichard HuangAI Infra EngineerUC San Diegoex-AWSRuirui ZhangSoftware EngineerUC Berkeleyex-AWSAce HaoAI Research EngineerUC Berkeleyex-AWSJacqueline ZhuChief of StaffHarvard UniversityTom LuAI ResearcherCambridge UniversityEx-QuantChristie-Carol BeauchampMobile Engineerex-MetaRobert IrvineAI ResearcherCambridge UniversityAlexandre DaoudBackend AI EngineerCambridge UniversityVineet MudupalliAI EngineerCambridge Universityex-QuantZongyi LiuAI ResearcherKaggle Masterex-Quant