ML Scientist - Japanese Multimodal
About the job
- Hands-on experience post-training modern language or multimodal models.
- Strong understanding of machine learning fundamentals and current post-training and reinforcement learning methods.
- Solid engineering skills and proficiency with the open-source machine learning ecosystem.
- Experience designing and running rigorous experiments, including baselines, ablations, and systematic error analysis.
- Experience building, curating, or assessing training and evaluation data at a meaningful scale.
- Ability to turn research ideas into reliable implementations and measurable model improvements.
- Proficiency in English, including the ability to collaborate on complex technical work with global teams.
- Experience leveraging agents to amplify your own work.
- Design and execute post-training strategies for language and multimodal models, including supervised fine-tuning, preference optimization, reinforcement learning, and distillation.
- Build and curate high-quality training data using human, synthetic, and model-generated signals, with particular attention to Japanese-language and domain-specific capabilities.
- Develop evaluations that expose meaningful capability and reliability gaps across Japanese and global use cases.
- Conduct systematic error analysis and use the results to improve data mixtures, objectives, training methods, and model behavior.
- Run controlled experiments and ablations, interpret results, and communicate clear recommendations.
- Develop reliable, scalable training and evaluation pipelines in collaboration with model infrastructure teams.
- Contribute methods, tooling, datasets, and findings that accelerate post-training work across Liquid AI.
- Experience with preference optimization or reinforcement learning methods for foundation models.
- Experience post-training multimodal models involving text, vision, or audio.
- Experience developing synthetic data pipelines, reward models, verifiers, or model-based evaluations.
- A track record of producing useful research artifacts, such as strong models, datasets, open-source systems, or technical reports.
- Reading proficiency in Japanese.
About the company
Liquid AI builds and deploys foundation models based on liquid neural networks, focusing on efficient, on-device AI. It develops Liquid Foundation Models (LFMs), a family of generative AI models designed to be smaller and more computation-efficient than typical large language models, enabling deployment on edge devices with lower latency, better privacy, and reduced infrastructure costs. The approach includes end-to-end AI expertise and customizable architectures for enterprises that require real-time performance and private processing. Compared to traditional AI providers, Liquid AI emphasizes edge-ready, adaptable models that run efficiently on constrained hardware, and it targets enterprise-grade, private, reliable AI solutions. The company’s goal is to enable real-time, on-device AI at scale for businesses by offering compact, capable foundation models and the tools to customize them for specific applications.
Company Size
51-200
Company Stage
Series A
Total Funding
$287.5M
Headquarters
Brookline, Massachusetts
Founded
2023
Get referred to Liquid AI
See people who can refer or advise you
Simplify's Take
What believers are saying
- Qualcomm partnership on Sep. 23, 2026 puts Liquid Context on Snapdragon Hexagon NPUs.
- MacPaw's Aug. 5, 2026 deal expands Liquid AI into millions of Mac users.
- DSpark on Aug. 20, 2026 cut H100 inference up to 3.18x without quality loss.
What critics are saying
- Mercedes-Benz, Qualcomm, and MacPaw create partner concentration before 2027 revenue diversification.
- Apple's on-device AI stack and OpenAI, Google commoditize local assistants on Macs and phones.
- The $10 million LFM license ceiling blocks larger customers, capping enterprise expansion.
What makes Liquid AI unique
- MIT spinout Liquid AI ships device-native LFMs for phones, cars, and NPUs.
- Pipette benchmarks full deployment stacks, exposing real latency across model, runtime, and hardware.
- d1 returns calibrated decisions with zero output tokens, not chatty text.
Help us improve and share your feedback! Did you find this helpful?
Benefits
Remote Work Options
Flexible Work Hours
Growth & Insights and Company News
6 month growth
3%1 year growth
0%2 year growth
4%Liquid AI releases open d1 models for multimodal edge decisions. Liquid AI released d1-3B for text and images and an experimental 600M model that pairs text with images or audio. By Ryan Merket - Published Oct 7th, 2026, 12:32pm CT - 4 MIN READ - 4 SOURCES - VIA Hugging Face Newsroom Listen to article. Liquid AI, co-founded by CEO Ramin Hasani, released two open-weight models in its d1 decision family on October 7th. The Hugging Face release introduces d1-3B, which accepts text and images, and experimental d1-omni-600M, which accepts text paired with either images or audio. Developers can download and run the models themselves, or try demos in Liquid AI's System One Arcade. Hasani has pursued the idea of building models that work within the limits of real devices since Liquid AI's founding in 2023. Before co-founding Liquid AI, Hasani was a CSAIL postdoctoral associate and a principal AI and machine-learning scientist at Vanguard. His research includes decision-making in complex systems, relevant to models designed to return a choice or score instead of composing an answer in prose. Liquid AI introduced an API-hosted d1 model in a company post on October 5th and said it planned to release open weights for upcoming models. Two days later, the Hugging Face post offered d1-3B and d1-omni-600M as open-weight downloads. Liquid AI is making separate decision-model checkpoints available for local deployment alongside its API. A model built to return a decision. Generative models produce tokens, even for narrow tasks such as sorting a support ticket or deciding whether an image shows a defect. Liquid AI's decision models instead process a state in a single forward pass and return structured answers. The release describes yes-or-no decisions, label selection and rubric-based scores as supported outputs. That format could suit routing, inspection and classification tasks where software needs an answer it can act on, rather than a paragraph it must parse. The two models use different foundations. d1-3B is built on Liquid AI's LFM2.5-VL-3B vision-language model. d1-omni-600M uses the LFM2.5-Encoder-350M bidirectional encoder, with added vision and audio encoders. Liquid AI labels the smaller model experimental and says it remains under development; it did not report speed numbers for it. The new downloads let developers evaluate local deployment, while the October 5th release described an API-hosted model. Liquid AI's announcement says both new models are available on Hugging Face and includes code showing d1-3B loaded through Transformers. It also instructs users to enable trust_remote_code=True. The release calls the models open-weight but does not state a license. Both Hugging Face model cards list LFM 1.0; teams evaluating deployment should review those terms for their intended use. Company-reported benchmarks and latency. In its Hugging Face release, Liquid AI reports a Decision Index 0.2.1 score of 48.57 for d1-3B, above the listed 4B and 9B models and Decider 35B-A3B at 47.11. For a separate comparison across seven public datasets, Liquid AI reports average scores of 82.9 for d1-3B and 78.4 for d1-omni-600M. Those figures exceed the listed Decider 4B and 2B averages, respectively. The October 7th announcement's benchmark tables do not include vision or audio scores. The d1-3B model card separately reports an average score of 74.1 across 11 public image benchmarks, compared with 73.9 for its LFM2.5-VL-3B base model. Liquid AI says the Decision Index v0.3 vision split is private and that audio decision benchmarks remain an open problem. The latency figures are also company-reported. In the same release, Liquid AI reports that d1-3B answered one question in 16 milliseconds on a Jetson AGX Thor, 26 milliseconds on a Jetson AGX Orin and 50 milliseconds on a Jetson Orin Nano. On the Thor, the reported time for three questions was 20 milliseconds. These results are specific to the tested devices and workloads; they do not establish performance across other hardware or tasks. Liquid AI's d1 models extend its device strategy to structured decisions and let developers run models outside a hosted API. Whether their speed and quality hold up in production tasks remains to be tested, particularly for the experimental audio-capable model. Liquid AI said its $250 million Series A, announced in December 2024 and led by AMD, would support model development and edge and on-premise product readiness, according to its funding announcement. The d1 release applies that stated goal to compact models that return machine-readable decisions locally. A separate Hugging Face benchmark tested d1 on September 30th across 400 cases and 2,000 decisions, reporting 0.742 accuracy; it did not directly replicate Liquid AI's seven-dataset comparison. RuntimeWire, Inc. get virtually zero search traffic. Please consider sharing this story. Signal check Noise Signal Where does this story land? Drag the dial. The startup Native device foundation models Liquid AI builds foundation models designed to run natively on devices outside of data centers, optimized for low latency, privacy, and real-world hardware constraints. Their LFM family spans compact to larger models and modalities, including LFM2.5-1.2B-Thinking (on-device reasoning under 1GB), LFM2-8B-A1B (efficient on-device mixture-of-experts), LFM2-24B-A2B (scaled-up LFM2 architecture), LFM2-VL series (efficient vision-language, including a 3B and a 450M model for structured visual intelligence), LFM2-Audio (end-to-end audio foundation model), and LFM2-ColBERT-350M for embeddings. They showcase tool-calling agents on consumer hardware and emphasize local, fast, private applications like meeting summarization. The company offers developer tooling such as LEAP and Liquid Apollo to make building with on-device AI easier, has documented use of Meta's ExecuTorch for efficient deployment, and highlights support on laptops with AMD Ryzen power. Notable collaborations and announcements include a partnership with Mercedes-Benz to scale embedded in-car intelligence, a multi-year partnership with Shopify to bring sub-20ms foundation models to core commerce experiences, demonstrations with AMD and Robotec.ai for embedded autonomy, and a $250M raise to scale capable and efficient general-purpose AI. A public playground allows users to try Liquid models directly. d1-omni-600M. Hugging Face model repository LiquidAI/d1-omni-600M, listed under the image-text-to-text task. This listing describes a public repository, not an inference endpoint. d1-3B. LFM2.5-Encoder-350M. Hugging Face model repository LiquidAI/LFM2.5-Encoder-350M, listed under the fill-mask task. This listing describes a public repository, not an inference endpoint. Hugging Face model repository LiquidAI/d1-3B, listed under the image-text-to-text task. This listing describes a public repository, not an inference endpoint. Hugging Face model repository Mapika/decider-4b, listed under the text-classification task. This listing describes a public repository, not an inference endpoint.
Multi-harness RL improves a small model across four agent interfaces. Federico Ulfo October 2, 2026 Researchers from Hugging Face and Liquid AI have released an open framework for training small language models inside multiple coding-agent harnesses. An agent harness controls the tools, context and execution loop around a model; changing it can substantially change how well the same weights perform. The framework combines OpenEnv, Harbor and TRL. A capture proxy records the exact tokens and generation probabilities needed for reinforcement learning while Claude Code, Codex, OpenCode and Mini-SWE-Agent run without changes to their code. On held-out SmolDataEnvs data-analysis tasks, multi-harness training raised LFM2.5-2.6B's average pass@1 from 42.2% to 54.2%, with improvements across all four harnesses. It also reduced tool calls by 31% on tasks that both the trained model and the baseline solved. A small efficiency bonus rewarded correct solutions that used fewer calls. Training only in OpenCode reached 52.3% overall and performed best within OpenCode. The 1.9-point overall gap between the two training approaches was within evaluation noise. The clearer advantage of multi-harness training was its broader efficiency gains and stronger results under Claude Code and Codex. Supervised fine-tuning on successful rollouts produced smaller accuracy gains. The work offers a practical way to train open models for several agent interfaces at once. These results come from a small model on data-analysis tasks, so they do not establish the same gains on large-scale software-engineering benchmarks or harnesses absent from training.
Liquid AI releases d1: A decision model that returns calibrated probabilities with zero output tokens. Reading Time: 14 mins read Liquid AI has released d1, a decision model built for structured choices instead of text generation. You give it context and a set of typed questions. It returns calibrated probabilities across a fixed set of outcomes in a single call, with zero generated tokens. The target is the work many teams still send to general LLMs: classification, ticket routing, scoring, moderation, reranking and LLM-as-judge checks. Is it deployable? Yes, today, as a hosted API. d1 runs on the Liquid API under the model name d1:free. Liquid's model library lists it as API only and not trainable, so there are no GGUF, MLX or ONNX weights to self-host. You may also like. What is a decision model? A decision model evaluates a situation and returns a typed answer from options you define before the call. It does not write text. In every response, usage.output_tokens is 0. Liquid AI's migration guide gives a simple rule: if the answer is one of N known options, use a decision model. If the model must compose a new string, keep your LLM. The 3 primitives: Noul, Choice and Score. * Noul is a yes/no question that returns a probability between 0 and 1. In Liquid's example, 'Is this message a complaint?' returned 0.999. * Choice picks one option from a named set. It returns the top pick, the full distribution and a confidence value. A double-charge ticket scored 0.9997 on 'billing.' * Score rates input on an ordered rubric and returns a probability-weighted position. Levels are indexed from 0, so a 4-level urgency rubric spans 0 to 3. A production outage scored 2.9995. You can mix all 3 types in one request. The model evaluates every question against the same state in one round trip. How a d1 API call works. Each request has 3 parts: the model, the state (plain text or a JSON object) and the questions. Calls go to POST https://api.liquid.ai/decisions/v1/systemone. Keys come from console.liquid.ai and start with liquid_. The clients are TypeSafe AI's typesafe-sdk for Python and @typesafe-ai/sdk for TypeScript. from typesafe_sdk import TypeSafeClient, Noul client = TypeSafeClient(api_key=os.environ["LIQUID_API_KEY"], base_url="https://api.liquid.ai") result = client.system_one( model="d1:free", state="I have been waiting over three weeks for my order...", questions={"is_complaint": Noul( instructions="Is this message a complaint from the customer?")},) print(result.answers["is_complaint"].noul) # 0.999 Why move LLM classification calls to d1. The migration guide lists the concrete differences against an LLM with structured output: * No billed output tokens: An LLM bills output even for a one-word label. * Predictable latency: There is no decoding loop that grows with output length. * No schema errors: Answers always match the question type, so malformed JSON and retries go away. * Usable uncertainty: You get calibrated probabilities instead of a self-reported number. * Fewer round trips: 3 sequential classification calls become 1. Probabilities make thresholds practical. Liquid's moderation example blocks above 0.8, allows below 0.2 and sends the middle band to human review. Its routing example falls back to the most capable model tier when router confidence drops below 0.5. Liquid also says repeated evaluations of the same input are more consistent, which reduces verdict flips. Keep an LLM for summarization, drafting, multi-turn chat, code generation and complex multi-step reasoning. Demo: road decider. Liquid's road-decider cookbook is a pixel-art survival racer. d1 uses a Choice question to pick left, center or right on every decision tick, about 2 to 5 times per second depending on game speed. The app is vanilla JavaScript on Node.js 18+, with a Vite proxy that keeps the API key server-side. A "Jev vs d1" mode races d1 against TypeSafe's typesafe/jev-1.13 through OpenRouter. The most useful lesson is about state design. Per-lane summaries with distance to the first obstacle produced more confident decisions than a raw grid of the road. Interactive explainer: d1 step by step. '+k+"+obj[k]+' ';el.appendChild(d);});setTimeout(function{var f=el.querySelectorAll('.fill');keys.forEach(function(k,i){f[i].style.width=Math.max(obj[k]*100,0.6)+'%'});},60);post;} function pills(el,list,cb){el.innerHTML=";list.forEach(function(x,i){var b=document.createElement('span');b.className="pill"+(i?":' on');b.textContent=x.name;b.onclick=function{el.querySelectorAll('.pill').forEach(function(p){p.classList.remove('on')});b.classList.add('on');cb(x)};el.appendChild(b)});cb(list[0]);} /* tabs */ var runners=; R.querySelectorAll('.tab').forEach(function(t){t.onclick=function{R.querySelectorAll('.tab').forEach(function(x){x.classList.remove('on')});R.querySelectorAll('.pane').forEach(function(x){x.classList.remove('on')});t.classList.add('on');$(t.dataset.p).classList.add('on');if(runners[t.dataset.p])runners[t.dataset.p];setTimeout(post,50)}}); /* p1 race */ var bill={billing:0.9997,account:0.0002,returns:0.00005,shipping:0.00003,technical:0.00002}; var toks=['{"','department','":',' "','billing','"}'];var timer; function race{clearInterval(timer);$('llmTok').innerHTML=";$('llmCnt').textContent="0";$('llmRes').textContent="...";$('d1Res').textContent="...";$('d1Bars').innerHTML=";$('llmParse').classList.remove('run');var c=$('d1Core');c.classList.remove('run');void c.offsetWidth;c.classList.add('run'); setTimeout(function{bars($('d1Bars'),bill,Object.keys(bill));$('d1Res').textContent="billing";},450); var i=0;timer=setInterval(function{if(i 0.999'},{name:'Harmful content',p:0.99,q:'state: "You\'re an absolute idiot... I\'m going to find out where you work."\nnoul: "Does this user message contain harmful, threatening, or abusive content?"\n=> 0.99'}]; function noulUpd{var hi=+$('hi').value,lo=+$('lo').value;$('hiV').textContent=hi.toFixed(2);$('loV').textContent=lo.toFixed(2);$('wiV').textContent=np.toFixed(3); R.querySelector('.gauge').style.setProperty('background','linear-gradient(90deg,#22c55e 0%,#22c55e '+lo*100+'%,#f59e0b '+lo*100+'%,#f59e0b '+hi*100+'%,#ef4444 '+hi*100+'%)','important'); $('needle').style.left="calc("+np*100+'% - 1.5px)';$('needleV').textContent=np.toFixed(3);var a=$('noulAct'); if(np>hi){a.textContent="block";a.style.background='#fde2e2';a.style.color="#b42318"}else if(np'+i+";$('scTicks').innerHTML=t;$('scNeedle').style.left="-1.5px";$('scV').textContent="0"; setTimeout(function{$('scNeedle').style.left="calc("+(x.s/mx*100)+'% - 1.5px)';$('scV').textContent=x.s},80);bars($('scBars'),x.p,Object.keys(x.p));$('scRes').textContent=x.s;$('scConf').textContent=x.c; $('scAct').textContent=x.n===4?(x.s>=2.5?'page on-call':x.s>=1.5?'escalate':'standard queue'):'escalate (SLA 4h)';})} runners.p4=scInit; /* p5 road */ var L=['left','center','right'],ROWS=6,ob=,cur=1;ob['0-1']=1;ob['2-3']=1; function road{var h="";for(var r=ROWS-1;r>=0;r-){for(var l=0;l'+(ob[k]?'traffic':'row '+(r+1))+'
Liquid AI, in collaboration with Qualcomm Technologies, brings personal AI Context to devices powered by Snapdragon. Liquid AI, at Snapdragon Summit 2026, announced that Liquid Context, an on-device context layer, is now optimized for Snapdragon(R) processors, specifically using the Qualcomm(R) Hexagon(TM) NPU. In collaboration with Qualcomm Technologies, Liquid Context built a persistent understanding of the user from signals from devices with Snapdragon processors and made relevant context available to AI agents, whether they ran on the device, in the cloud or across both. Device manufacturers now have a foundation for more proactive, personalized and trustworthy AI experiences, alongside Liquid Agent, Liquid AI's efficient embedded agent. Personal Context That Makes Any Agent More Useful Useful agents need to understand what matters to the user and when help is needed. With the user's permission, Liquid Context learns from device signals, builds an understanding of routines, preferences and needs, and keeps that understanding current. It provides a shared context layer between the device and the user's chosen agents, including third-party agents and Liquid Agent. Context supplies the understanding; agents use it to reason, suggest next steps and take action with permission. Personal context is built and maintained locally, with relevant context made available to connected agents according to the user's permissions. "Personal AI starts with understanding how you live and what you need, when you need it," said Liquid AI CEO and co-founder Ramin Hasani. "Liquid Context builds that understanding on your device so the agents you choose can offer more relevant help and anticipate your needs. The Hexagon NPU makes this continuous, local intelligence practical. Together, we are bringing personal context to the devices people rely on every day." Efficient NPU execution is central to keeping this understanding current. Liquid Context is designed to run in the background, turning permitted device signals into useful context without requiring a cloud model to process every update. Agents can then use that context to recognize when help would be valuable and respond in a way that reflects the user's priorities. Illustrative experiences enabled by Liquid Context include: * Adapting to a family emergency: a message arrives that the user's child is sick and needs an early school pickup. Liquid Context supplies relevant family context and the user's priorities. A connected agent checks the calendar, identifies which meetings can move while preserving an important commitment, and drafts rescheduling emails for the user to approve. * Turning a conference day into a personal recap: after a day of networking and keynotes, Liquid Context supplies the user's preference for sharing a LinkedIn recap and their writing style, alongside relevant context from the day. A connected agent selects photos the user has permitted it to access, drafts a post in their voice and queues it for review. * Carrying context from a run into the car: the user leaves their phone at home and runs with a smartwatch. The watch captures local context about the run, including a new personal record. When the user returns to the car, that context syncs locally with permission. The car's agent congratulates them, cools the cabin according to their preferences and suggests a stop for a recovery smoothie. Sep 24, 2026 Prev Next 1 of 43,684 Bringing Liquid Context to Snapdragon Platforms Qualcomm Technologies and Liquid AI are exploring additional opportunities to work together to further enhance the performance of Liquid Context with Snapdragon platforms, so OEMs can offer personal context as a built-in device capability, making it easier for users to enable Liquid Context. Compatible embedded, cloud and hybrid agents could even use that context to deliver proactive, personalized assistance. For OEMs choosing a platform for agentic experiences, the optimization brings together Hexagon NPU hardware built for AI workloads and Liquid Context software that makes device signals useful to agents. It gives manufacturers a path to add personal context while supporting their choice of agents and services. "Qualcomm Technologies is at the forefront of advancing AI across devices, edge solutions and datacenters, driving the innovations needed to deliver increasingly intelligent, personalized experiences for consumers around the world. By combining our industry-leading Snapdragon platforms with innovative AI technologies from companies like Liquid AI, we're helping accelerate the next generation of Agentic AI experiences." - Durga Malladi, EVP & GM, Technology Planning, Edge Solutions, and Data Center, Qualcomm Technologies, Inc. Hexagon NPU: Built for Context and Agents Continuous context processing and responsive agent execution both require efficient hardware. Liquid AI has optimized its Liquid Context memory layer and, separately, the LFM2.5-2.6B agentic model that powers Liquid Agent to take advantage of processing, memory access and low latency of Snapdragon processors. The next-generation Hexagon NPU combines scalar, vector and matrix processing with dedicated transformer hardware to accelerate these workloads. This pairing can support ongoing context updates and capable embedded agents within the power constraints of everyday devices. Liquid Agent: An Efficient Embedded Agent for OEMs OEMs building embedded or hybrid agents can also work with Liquid AI to evaluate Liquid Agent, powered by its state-of-the-art, efficient LFM2.5-2.6B agentic model and optimized for execution on the Hexagon NPU. Liquid Agent can be tailored to a manufacturer's hardware, services, interface and brand, and can consume Liquid Context to deliver personalized, proactive assistance. Liquid Context provides the shared context layer for that vision, preserving relevant understanding and task state with permission so different agents can pick up where the user left off. OEMs can use the Hexagon NPU and Liquid Context as a foundation for their chosen agents, and evaluate Liquid Agent when they need an efficient embedded or hybrid agent of their own.
Liquid AI, in collaboration with Qualcomm Technologies, brings personal AI Context to devices powered by Snapdragon. Business Wire Published Sep 23, 2026 MAUI, Hawaii - Liquid AI today, at Snapdragon Summit 2026, announced that Liquid Context, an on-device context layer, is now optimized for Snapdragon(R) processors, specifically using the Qualcomm(R) Hexagon(TM) NPU. In collaboration with Qualcomm Technologies, Liquid Context built a persistent understanding of the user from signals from devices with Snapdragon processors and made relevant context available to AI agents, whether they ran on the device, in the cloud or across both. Device manufacturers now have a foundation for more proactive, personalized and trustworthy AI experiences, alongside Liquid Agent, Liquid AI's efficient embedded agent. Personal Context That Makes Any Agent More Useful Top Stories Interested in more newsletters? Browse here. Useful agents need to understand what matters to the user and when help is needed. With the user's permission, Liquid Context learns from device signals, builds an understanding of routines, preferences and needs, and keeps that understanding current. It provides a shared context layer between the device and the user's chosen agents, including third-party agents and Liquid Agent. Context supplies the understanding; agents use it to reason, suggest next steps and take action with permission. Personal context is built and maintained locally, with relevant context made available to connected agents according to the user's permissions. "Personal AI starts with understanding how you live and what you need, when you need it," said Liquid AI CEO and co-founder Ramin Hasani. "Liquid Context builds that understanding on your device so the agents you choose can offer more relevant help and anticipate your needs. The Hexagon NPU makes this continuous, local intelligence practical. Together, we are bringing personal context to the devices people rely on every day." Efficient NPU execution is central to keeping this understanding current. Liquid Context is designed to run in the background, turning permitted device signals into useful context without requiring a cloud model to process every update. Agents can then use that context to recognize when help would be valuable and respond in a way that reflects the user's priorities. Illustrative experiences enabled by Liquid Context include: * Adapting to a family emergency: a message arrives that the user's child is sick and needs an early school pickup. Liquid Context supplies relevant family context and the user's priorities. A connected agent checks the calendar, identifies which meetings can move while preserving an important commitment, and drafts rescheduling emails for the user to approve. * Turning a conference day into a personal recap: after a day of networking and keynotes, Liquid Context supplies the user's preference for sharing a LinkedIn recap and their writing style, alongside relevant context from the day. A connected agent selects photos the user has permitted it to access, drafts a post in their voice and queues it for review. * Carrying context from a run into the car: the user leaves their phone at home and runs with a smartwatch. The watch captures local context about the run, including a new personal record. When the user returns to the car, that context syncs locally with permission. The car's agent congratulates them, cools the cabin according to their preferences and suggests a stop for a recovery smoothie. Bringing Liquid Context to Snapdragon Platforms Qualcomm Technologies and Liquid AI are exploring additional opportunities to work together to further enhance the performance of Liquid Context with Snapdragon platforms, so OEMs can offer personal context as a built-in device capability, making it easier for users to enable Liquid Context. Compatible embedded, cloud and hybrid agents could even use that context to deliver proactive, personalized assistance. For OEMs choosing a platform for agentic experiences, the optimization brings together Hexagon NPU hardware built for AI workloads and Liquid Context software that makes device signals useful to agents. It gives manufacturers a path to add personal context while supporting their choice of agents and services. "Qualcomm Technologies is at the forefront of advancing AI across devices, edge solutions and datacenters, driving the innovations needed to deliver increasingly intelligent, personalized experiences for consumers around the world. By combining our industry-leading Snapdragon platforms with innovative AI technologies from companies like Liquid AI, we're helping accelerate the next generation of Agentic AI experiences." - Durga Malladi, EVP & GM, Technology Planning, Edge Solutions, and Data Center, Qualcomm Technologies, Inc. Hexagon NPU: Built for Context and Agents Continuous context processing and responsive agent execution both require efficient hardware. Liquid AI has optimized its Liquid Context memory layer and, separately, the LFM2.5-2.6B agentic model that powers Liquid Agent to take advantage of processing, memory access and low latency of Snapdragon processors. The next-generation Hexagon NPU combines scalar, vector and matrix processing with dedicated transformer hardware to accelerate these workloads. This pairing can support ongoing context updates and capable embedded agents within the power constraints of everyday devices. Liquid Agent: An Efficient Embedded Agent for OEMs OEMs building embedded or hybrid agents can also work with Liquid AI to evaluate Liquid Agent, powered by its state-of-the-art, efficient LFM2.5-2.6B agentic model and optimized for execution on the Hexagon NPU. Liquid Agent can be tailored to a manufacturer's hardware, services, interface and brand, and can consume Liquid Context to deliver personalized, proactive assistance. Liquid Context provides the shared context layer for that vision, preserving relevant understanding and task state with permission so different agents can pick up where the user left off. OEMs can use the Hexagon NPU and Liquid Context as a foundation for their chosen agents, and evaluate Liquid Agent when they need an efficient embedded or hybrid agent of their own. This advertisement has not loaded yet. This advertisement has not loaded yet, but your article continues below. About Liquid AI Liquid is a device-native foundation model company. We bring advanced AI to all processors outside of data centers and are building the software layer of physical AI. Any device (e.g., phone, laptop, car, airplane, robot, wearables, etc.) with a processor gets powered to run and tune advanced liquid foundation models to deliver reliable and efficient intelligence in the real world. Liquid's open-weight flagship Foundation Models (LFMs) establish the Pareto frontier of efficient intelligence on devices, enabling seamless deployment of advanced AI directly on devices where latency, privacy, and performance matter most. Learn more at liquid.ai. Snapdragon and Qualcomm branded products are products of Qualcomm Technologies, Inc. and/or its subsidiaries. Qualcomm, Snapdragon and Hexagon are trademarks or registered trademarks of Qualcomm Incorporated. View source version on businesswire.com: Media contact Sarah Mulligan | Quick Picks