Back
  • Product
  • Solutions
  • Resources
  • Open Source
  • Docs
  • Blog
  • Company
Request a demo
Sign up
    • Inference Products

      • Shared Endpoints

        Access frontier models via an API

      • Dedicated Endpoints

        Mission critical reliability

      • Custom models

        Your model, peak performance

    • Deployment Options

      • Our Cloud

        Fully managed, pay by usage

      • Your Cloud

        Modular stack in your VPC

      • Pricing

        Flexible plans for every team

    • Models

      FLUX.2 Klein 9B logo

      FLUX.2 Klein 9B

      LTX-2.3 logo

      LTX-2.3

      MiniMax M3 logo

      MiniMax M3

      Kimi K2.7 Code logo

      Kimi K2.7 Code

      Inkling logo

      Inkling

      View All

      • Text to audio

        Turn text into natural speech

      • Image generation

        Generate images from text prompts

      • Code generation

        Generate production-ready code

      • Video generation

        Generate video from text + image

      • Agentic

        Deploy AI agents anywhere

      • Custom Models

        Kernel-level model control

      • Case Studies

        Proven results from real customers

    • Models

      FLUX.2 Klein 9B logo

      FLUX.2 Klein 9B

      LTX-2.3 logo

      LTX-2.3

      MiniMax M3 logo

      MiniMax M3

      Kimi K2.7 Code logo

      Kimi K2.7 Code

      Inkling logo

      Inkling

      View All

      • MAX Framework

        GenAI native modeling & serving

      • Mojo Language

        The systems language for the AI era

      • Self-Hosted

        MAX+Mojo self-hosted by you

      • Community

        Build the future of AI together

      • Agentic Skills

        Official AI agent skills from Modular

    • Docs

      Deploy GenAI models, our cloud or yours

    • Model Library

      Latest supported open models

    • Mojo Docs

      The systems language for the AI era

    • LLM Inference Handbook

      Learn how inference works, from basics to production

    • About

      Build AI for anyone, anywhere.

    • Careers

      👋  We’re currently hiring!

    • Culture

      What we believe

    • Contact Us

      Request a demo

Request a demo
Sign up
close

Modular Community License

‍Last Modified: August 18, 2026

These Modular MAX Community License terms (these "Terms") are entered into by and between Modular Inc ("Modular") and the individual or entity agreeing to these Terms (as part of "You" and “Your” as defined in the Terms). These Terms govern Your use of the MAX software development kits (“MAX”).  

1. LICENSE GRANT.

1.1 MAX SDK Limited Source License. Conditioned upon Your compliance with these Terms (including the Attribution and Additional Requirements as provided below), Modular hereby grants You a perpetual, worldwide, non-exclusive, no-charge, royalty-free license to (a) reproduce and prepare Derivative Works; (b) publicly display and publicly perform Derivative Works in object code form; (c) sublicense and distribute Derivative Works in object code form; and/or (d) use accompanying MAX Documentation provided by Modular (“Documentation”) solely internally for the purposes of (a) and (b) above.

1.1.1 Derivative Works. "Derivative Work" means any work, whether in source or object form, that is based on (or derived from) MAX and for which the editorial revisions, annotations, elaborations, or other modifications that represent, as a whole, an original work of authorship;

1.1.2 Downstream License for Derivative Works (Object Code Only). The terms under which You distribute Derivative Works must be consistent with the terms of this Agreement, including (without limitation) terms relating to the license grant and license restrictions and protection of Modular’s intellectual property rights;

1.1.3. Display of Terms. These Terms shall be conspicuously displayed (such as via a “Notice” text file distributed in the top-level directory as part of such Derivative Works) in each copy of a Derivative Work;

1.1.4 Attribution. You must prominently display the following notice alongside the distribution of any Derivative Works:

“MAX is licensed by Modular Inc under the Modular MAX Community License and has been modified by [You]. Copyright Modular Inc IN NO EVENT SHALL MODULAR INC BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH USE OF MAX OR ANY DERIVATIVE WORK THEREOF.”

1.2 MAX Application Development License.  In addition to Your license to use the MAX SDK pursuant to Section 1.1, certain portions of the MAX SDK may be redistributed as part of an application You develop (“Application”) as a result of your authorized use of the MAX SDK.  Conditioned upon Your compliance with these Terms, Modular hereby grants to You the limited right to redistribute certain components of MAX as part of Your Applications, in object code form, subject to the conditions and limitations set forth below and the Application Distribution Requirements. You shall only redistribute the Redistributable Components as expressly permitted in these Terms and in the Documentation, in accordance with the Application Distribution Requirements.

“Application Distribution Requirements” means each of the following requirements:

  1. The Application must have material additional functionality, beyond the included portions of MAX;
  2. The terms under which You distribute an Application must be consistent with the terms of these Terms, including (without limitation) terms relating to the license grant and license restrictions and protection of Modular’s intellectual property rights;

1.3 Additional Requirements.

  1. You shall preserve and not obscure any proprietary notices, labels, and marks provided by Modular in MAX, Derivative Works of MAX, Redistributable Components, and Documentation;
  2. You agree to notify Modular in writing of any known or suspected violation of these Terms and to reasonably enforce the terms of a Derivative Work or an Application. If You receive MAX in original or modified form from any third party, You agree and accept that these Terms are still binding upon You.
  3. You shall not misrepresent or imply, through any means, that a Derivative Work is a product of Modular or have been endorsed, approved, or validated by Modular.
  4. You shall not use MAX, or any portion of it, as training data, fine-tuning data, or input to any AI system in order to produce software that reimplements or substitutes for MAX. is prohibited, whether or not it contains literal copies of MAX. This does not restrict using AI tools to read, analyze, improve, or explain MAX, or to develop Your own software that runs on or interoperates with MAX.

1.4 Separately Licensed Components. Portions of MAX have been made available under the Apache License, Version 2.0 with LLVM Exception or under other separate license terms, as identified in the LICENSE file or license header accompanying each such component. Those terms govern those components, and control over these Terms in the event of any conflict. Nothing in these Terms revokes, narrows, conditions or otherwise modifies any right granted under any such separate license.

1.5 Patent License.  Conditioned upon Your compliance with these Terms, Modular hereby grants to You a perpetual, worldwide, non-exclusive, no-charge, royalty-free, irrevocable (except as stated in this section) patent license to make, have made, use, offer to sell, sell, import, and otherwise transfer MAX as delivered to You by Modular. If You institute patent litigation against any entity (including a cross-claim or counterclaim in a lawsuit) alleging that MAX constitutes direct or contributory patent infringement, then any patent licenses granted to You under these Terms shall terminate as of the date such litigation is filed.


2. USAGE DATA; PERSONAL INFORMATION.

2.1 Usage Data. You acknowledge and understand that Your use of MAX may be subject to the collection of certain telemetry, usage, and other data which captures Your interactions with and use of MAX (collectively, the "Usage Data"). You agree that Modular and its affiliates may freely use, modify, copy, and/or distribute the Usage Data for any purpose, including but not limited to improving MAX or any other component or aspect of Modular's and/or its affiliates’ products. As between the parties, You acknowledge that Modular owns all right, title, and interest in and to the Usage Data and may use it for any legitimate purpose, including but not limited to the purposes of improving MAX or any other component or aspect of the Modular Platform.

2.2 Personal Information. Any collection, storage, or use of Your personal information is governed by Modular’s Privacy Policy (https://www.modular.com/legal/privacy).

‍

3. OWNERSHIP; FEEDBACK; TRADEMARK LICENSE.

3.1 Ownership. Modular retains all right, title, and interest in and to MAX and including but not limited to Redistributable Components, including all intellectual property rights therein. Other than as explicitly stated in these Terms, no other rights are granted to MAX.

3.2 Feedback. If You share any suggestions, recommendations, or changes to MAX or the Documentation, including without limitation, new features or functionality relating thereto, or any comments, questions, suggestions, or the like (collectively, “Feedback”), Modular and its affiliates are free to use such Feedback without any restriction, obligation, or compensation to You.

3.3 Trademark License. If You (a) are a legal entity, and (b) commercially deploy MAX as a hosted or managed AI training or inference service to other legal entities (“Authorized Product”), You shall use the Powered by Modular word mark and/or logo, or any other word mark and/or logo that may be designated by Modular (collectively, the “Marks”), in connection with the Authorized Product. You shall download the Marks from the website to be designated by Modular (the “Brand Portal”) and agree to any applicable terms of use (the “Brand Portal Terms of Use”). You shall adhere to all usage and marking requirements set forth in the brand guidelines for the Marks on the Brand Portal, which may be revised from time to time in Modular’s sole discretion (the “Brand Guidelines”). You shall submit all materials bearing the Marks to Modular via the Brand Portal or usage@modular.com and shall obtain Modular’s written approval prior to any use thereof. Modular may withhold its consent or cancel prior authorization for use of the Marks if the proposed usage violates the Brand Portal Terms of Use and/or Brand Guidelines. Upon termination of or cessation of deployment of the Authorized Product, Your obligations under this Section 3.3 shall terminate, and You shall immediately discontinue all use of the Marks and any logos, marks, or terms confusingly similar thereto. Nothing herein shall be construed as a grant of a license or any other rights with respect to Modular’s or its affiliates’ trademarks, word marks, service marks, logos, trade dress, or other branding.

‍

4. SUPPORT AND UPDATES. Modular is not obligated to provide any support, patches, fixes or updates for MAX. However, Modular may, at its sole discretion, provide updates, patches, fixes or support at any time.

‍

5. TERMINATION. You may terminate these Terms by ceasing all use of MAX and any Documentation. Modular may terminate these Terms upon Your material breach of any provision.

‍

6. WARRANTY DISCLAIMER. MAX AND ANY DOCUMENTATION ARE PROVIDED "AS IS" AND TO THE EXTENT PERMITTED BY APPLICABLE LAW, MODULAR MAKES NO OTHER WARRANTIES, EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO WARRANTIES OF MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE, OR NON-INFRINGEMENT. YOU ACKNOWLEDGE AND AGREE THAT THE USE OF MAX IS AT YOUR OWN RISK, AND MODULAR DOES NOT MAKE ANY REPRESENTATIONS OR WARRANTIES REGARDING THE ACCURACY, RELIABILITY, OR COMPLETENESS OF MAX, OR INFORMATION OBTAINED THROUGH MAX. MODULAR DOES NOT WARRANT OR GUARANTEE THAT MAX WILL MEET YOUR REQUIREMENTS OR EXPECTATIONS FOR ITS APPLICATIONS OR GENERALLY, OR THAT ANY ERRORS OR DEFECTS IN MAX WILL BE CORRECTED. FURTHERMORE, MODULAR DOES NOT WARRANT OR MAKE ANY REPRESENTATIONS REGARDING THE RESULTS THAT MAY BE OBTAINED FROM THE USE OF MAX OR THE SUITABILITY OF MAX FOR ANY PARTICULAR PURPOSE. MODULAR MAKES NO WARRANTIES OR REPRESENTATIONS THAT MAX WILL BE COMPATIBLE WITH YOUR SYSTEMS, SOFTWARE, OR APPLICATIONS, OR THAT MAX WILL BE SECURE FROM UNAUTHORIZED ACCESS, HACKING, OR OTHER POTENTIAL SECURITY THREATS. MODULAR DISCLAIMS ANY LIABILITY ARISING OUT OF OR IN CONNECTION WITH THE USE OF MAX. YOU ASSUME ALL RISKS AND RESPONSIBILITIES FOR USING MAX TO ACHIEVE YOUR INTENDED RESULTS AND FOR THE INSTALLATION, USE, AND RESULTS OBTAINED FROM MAX. ADDITIONALLY, VERSIONS OF MAX IDENTIFIED AS ‘PREVIEW VERSIONS’ MAY HAVE ADDITIONAL BUGS OR DESIGN FLAWS. MODULAR MAKES NO WARRANTIES ABOUT THE PREVIEW VERSIONS WHATSOEVER, AND YOUR USE OF MAX IN COMMERCIAL, PRODUCTION ENVIRONMENTS IN VIOLATION OF THIS AGREEMENT IS AT YOUR OWN RISK.

‍

7. INDEMNIFICATION. You shall defend, indemnify and hold harmless Modular, its affiliates and their respective officers, directors, employees, agents and representatives from any and all third party claims, damages, liabilities, costs and fees (including reasonable attorneys’ fees) arising from: (i) Your material breach of these Terms; or (ii) an allegation that any Application or Derivative Work violates, misappropriates, or infringes any third party intellectual, contractual or proprietary right.

‍

8. LIMITATION OF LIABILITY. TO THE EXTENT PERMITTED BY APPLICABLE LAW, IN NO EVENT SHALL MODULAR OR ITS LICENSORS BE RESPONSIBLE OR LIABLE WITH RESPECT TO ANY SUBJECT MATTER OF THIS AGREEMENT UNDER ANY CONTRACT, NEGLIGENCE, STRICT LIABILITY OR OTHER THEORY FOR: (I) LOSS OR INACCURACY OF DATA OR COST OF PROCUREMENT OF SUBSTITUTE GOODS, SERVICES, OR TECHNOLOGY; OR (II) ANY INDIRECT, INCIDENTAL, OR CONSEQUENTIAL DAMAGES INCLUDING, BUT NOT LIMITED TO LOSS OF REVENUES OR PROFITS; (III) ANY MATTER BEYOND ITS REASONABLE CONTROL; OR (IV) ANY AMOUNT IN THE AGGREGATE OVER $100. SOME JURISDICTIONS DO NOT ALLOW CERTAIN WARRANTY DISCLAIMERS OR LIMITATIONS ON LIABILITY. ONLY DISCLAIMERS OR LIMITATIONS THAT ARE LAWFUL IN THE APPLICABLE JURISDICTION WILL APPLY TO YOU, AND MODULAR’S LIABILITY WILL BE LIMITED TO THE MAXIMUM EXTENT PERMITTED BY LAW.

‍

9. CHANGES TO THESE TERMS. The version of these Terms that accompany, or is displayed at the time of download or installation of, a given release of MAX governs Your use of such release. Modular may modify these Terms from time to time. Each version of these Terms will be identified by a Last Modified date (the “Effective Date”), posted at https://www.modular.com/legal/community.

Modified Terms shall apply to releases of MAX made available on or after the Effective Date. By downloading, installing, or using any release of MAX made available on or after the Effective Date, You accept those modified Terms. Notwithstanding the foregoing, you may continue to use a version of MAX under the terms of that license that applied when you downloaded that version unless that license was terminated for material breach under Section 5.

‍

Qualcomm + Modular ModCon 2026 External Communications FAQ

General Q&A

‍

Q: What did Modular announce at ModCon 2026?

A: The theme was unlocking compute through open source, open cloud, and open silicon. Modular Platform is production-ready today, serving billions of tokens per minute in enterprise deployments. There were six standout product announcements:

  1. Mojo 1.0 is fully open source under an Apache 2.0 license with LLVM exceptions. The compiler and all tooling are open.
  2. Modular Cloud is publicly available at console.modular.com and serving flagship customers, including MiniMax.
  3. Modular Platform now supports Qualcomm Cloud AI 100 and Dragonfly AI200 accelerators.
  4. Modular Platform now supports Google TPUs and AWS Trainium alongside CPUs and GPUs.
  5. MAX becomes source-available under a more open license, and Modular is working toward an industry alliance so partners can build the platform alongside Modular.
  6. Native Windows support for Mojo is on the horizon, with the Microsoft Windows team working alongside Modular.

‍

Q: What is Modular, and what is its relationship to Qualcomm now?

A: Modular builds an open, AI-native software platform that lets developers run AI across different kinds of hardware, from kernel to cloud, without being locked into one vendor. The same model runs across CPUs, GPUs, NPUs and custom ASICs without a rewrite for each accelerator. Build once, deploy anywhere, and spend less running AI at scale.

Qualcomm acquired Modular, Inc. Modular still operates as an open, vendor-neutral platform business. Chris Lattner, who founded Modular with Tim Davis, is EVP, AI Software and Platforms at Qualcomm and remains a co-founder of Modular.

‍

Q: Why does ModCon matter for Qualcomm?

A: Modular provides an open, heterogeneous, vendor-neutral software platform for AI infrastructure, and ModCon is our annual developer and ecosystem conference. It matters for Qualcomm for two reasons.

First, it showcases the broader software transformation Qualcomm is making as it becomes a more developer-first company – opening its technology to a wider ecosystem and making it easier for developers to build on Qualcomm platforms. Mojo OSS, MAX becoming source-available under a more open license are proof of that, and Modular Platform integrations with AWS Trainium and Google TPUs are proof of that.

Second, ModCon is where that strategy becomes tangible. For the first time, the Modular–Qualcomm combination shows up in technology customers can run today. Qualcomm Technologies’ data-center AI accelerators are now integrated into the Modular Platform, allowing software built for NVIDIA and AMD GPUs to target Qualcomm silicon (e.g., Dragonfly) through the same abstractions and developer experience.

‍

Q: Is Modular now a Qualcomm-only stack, and does this challenge NVIDIA and CUDA?

A: Modular is not Qualcomm-only. Joining Qualcomm does not narrow the goal, and the ModCon announcements are the evidence: on the same day Modular announced Qualcomm accelerator support, it announced support for Google TPUs and AWS Trainium. Modular Platform will keep supporting and optimizing for a broad range of hardware, including hardware that competes directly with Qualcomm. The opportunity is much bigger than any single vendor's roadmap, and a foundation only works if everyone can stand on it.

‍

Q: Does Modular really outperform hardware vendor software on their native hardware?

A: Yes. On production workloads, Modular outperforms existing vendor and open source stacks even on the vendor's own silicon. Serving models at scale takes more than matrix multiplication. It is prefill and decode balance, KV-cache management, batching, scheduling, memory movement, quantization, speculative decoding, multi-device execution and disaggregated inference. Modular rebuilt the stack from low-level kernels to cloud orchestration, so optimizations compound across layers.

‍

Q: What role does Modular play with edge devices?

A: Modular built the platform to span data center and edge, and support can extend there over time. MAX already runs on some edge-class devices as a byproduct of the data center work, including consumer NVIDIA and AMD accelerators, Apple GPUs on laptops and Snapdragon chips.

‍

Customer, Partner and Ecosystem Q&A

‍

‍
‍Q: What is Modular Cloud, and who is using it?

A: Modular Cloud is the full Modular stack delivered as a production service, publicly available at console.modular.com. It gives developers Modular's inference performance without the complexity of deploying, optimizing and operating models across heterogeneous infrastructure. Shared endpoints are OpenAI-compatible with pay-per-token pricing. Dedicated deployments run on Modular compute or the customer's own reserved instances, and customers can also run Modular Platform entirely in their own environment.

MiniMax is a flagship enterprise customer, running M3 on a dedicated deployment that serves production traffic at billions of tokens per minute. M3 is hard to serve: a 1M-token context window, native multimodality, and MiniMax Sparse Attention. Getting there meant implementing M3 natively in MAX and tuning specialized MSA kernels. Modular Cloud has also been carrying OpenRouter traffic for months under the name ModelRun, where it ranks at or near the top of the platform on latency and throughput. Artificial Analysis, an independent benchmarking firm, reports the same.

‍

Q: What does support for TPUs, Trainium and Qualcomm accelerators mean, and how mature is it?

A: Each one runs through the same Modular Platform, using the same modeling APIs, serving workflows, programming language and core abstractions as NVIDIA and AMD GPUs. Developers write a model once and bring it to entirely different hardware architectures without rebuilding the software stack around it.

Maturity varies by platform. Qualcomm AI100 support extends through Modular Cloud. TPU and Trainium can execute Gemma models through MAX, but aren’t yet integrated into Modular Cloud. Coverage across all these platforms will keep expanding, and we expect them to be production ready within months.

Notably, each of these platforms was brought up with roughly a tenth of the engineering effort normally needed to enable new AI hardware.

‍

‍Q: What is Modular opening up, and what is the industry alliance?

A: The Mojo language – including its standard library, compiler, and tooling – is now available in Modular’s GitHub repository under the Apache 2.0 license with LLVM exceptions. This continues the progressive opening of the stack that began with the Mojo standard library in 2024 and MAX kernels in 2025. We’ve accepted contributions to the Mojo standard library since 2024, and while we’re not yet taking contributions to the compiler and tooling, we expect to open those up to contributions over time.

We’re also significantly opening up MAX. Building on the portions of MAX already available under Apache 2.0, we’re making the remaining MAX source – including the compiler, runtime, and device APIs – available under a more permissive Community License. The new license removes usage restrictions and provides distribution rights, so developers, enterprises, hardware vendors, and partners can build on MAX, extend it for their own needs, and contribute improvements back to the platform.

Finally, Modular is working toward an industry alliance bringing together hardware vendors, model providers, cloud companies, and data-center operators. Participating partners will have a direct role in integrating MAX with their platforms, testing and optimizing those integrations, and helping shape the future roadmap of the Modular Platform. We expect to share more details on the alliance by the end of the year.

‍

Q: What does this mean for other hardware vendors?

A: It helps them. AI hardware innovation is accelerating, but great hardware only matters if developers can use it, and too much of that work is still gated behind software stacks written for one architecture. The industry needs a common foundation instead: write a model once and reach every accelerator, choose hardware on performance and economics rather than on which stack happens to support it, and let vendors compete on the merits of their silicon. Under Qualcomm, Modular can move faster on open development of Mojo and MAX, which shortens every vendor's path to market.

‍

Q: Where can people go to learn more?

A: Open source Mojo 1.0 is at mojolang.org and github.com/modular/modular. Modular Cloud is live at console.modular.com. ModCon information is at modular.com/modcon. Technical and contribution questions go to forum.modular.com.

‍

Modular MAX Community License Q&A

‍

Q: What prompted the change to the Community License?

A: From the beginning, we’ve believed that open software is the best foundation for a healthy, durable ecosystem. At the same time, Mojo and MAX needed room to incubate within Modular while we developed the core architecture and brought the products to maturity.

We’ve now reached the point where we believe both are ready to open up more broadly to the ecosystem. Mojo is moving to the Apache 2.0 license with LLVM Exceptions, while additional parts of MAX are becoming source-available under our new Community License. This gives developers, enterprises, hardware vendors, and partners much greater visibility into and ability to build on the platform, while giving us the framework we need to continue developing MAX as a sustainable, broadly adopted product.

‍

Q: What’s changed in the new Community License?

A: The old license capped free production use at eight accelerators outside x86, ARM and NVIDIA, and required you to ask us in writing before running on custom hardware we hadn't expressly supported. Both requirements are now gone. 

The new license also lets you build on MAX using the source that we are making available. You ship what you build in object code form and you can build and ship applications, as long as you follow the attribution and license requirements. Note that if you use MAX for an enterprise hosted or managed AI training or inference service, you must follow our trademark requirements.

Anything already open sourced stays that way. We are not relicensing it, and Section 1.4 of the license makes that explicit.

‍

Q: How can I use MAX with AI tools?

A: We publish the source for MAX because we want people to learn from it, improve it and build on it. Use Claude, Copilot, Cursor or anything else to read MAX, understand how it works, debug against it, write patches for it, or build your own software on top of it.

The one thing you can't do is point an AI at MAX in order to produce a replacement for MAX. This includes feeding our source to a model to generate a reimplementation, a port to another language, or a substitute runtime.

‍

Qualcomm branded products are products of Qualcomm Technologies, Inc. and/or its subsidiaries. Qualcomm, Qualcomm Dragonfly, and Hexagon are trademarks or registered trademarks of Qualcomm Incorporated.

‍

Latest from our blog:

New

ModCon 2026: Open source, open cloud, open silicon

Get the latest news,
announcements & updates:

Join our Newsletter
  • INFERENCE

    • Shared Endpoints
    • Dedicated Endpoints
    • Custom Models
  • DEPLOYMENTS

    • Our Cloud
    • Your Cloud
    • Pricing
  • Solutions

    Text to audio
    Image generation
    Code generation
    AI Agents
  • Open SOURCE

    • MAX
    • Mojo🔥
    • Agent Skills
  • Connect

    • Blog
    • Community
    • Report a security issue
  • Company

    • About Us
    • Culture
    • Careers
    • Request a demo

Copyright © 2026 Modular Inc, A Qualcomm Company

Do not sell or share my personal information

Terms, Privacy & Acceptable Use

Join our newsletter

Get all our latest news, announcements and updates delivered directly to your inbox. Unsubscribe at anytime.

Your data is handled in accordance with Modular's Privacy Policy.

Thanks for signing up to our newsletter! 🚀

Thank you,

Modular Sales Team

Oops! Something went wrong while submitting the form.

Return to page

Modular Champions

X/X

Owen Hilyard

I'm a PhD student at the University of New Hampshire in the Cloud Computing Lab, where I conduct research on hardware acceleration of networking and distributed systems reliability. I'm also a former component maintainer for DPDK, the Data Plane Development Kit, which is where I got my start looking "under the hood" at networking before resigning to work in my PhD. All of this means I like making computers go fast, and Mojo + MAX is a great place to combine my love of hardware, high performance software, and my interest in programming languages. I also act as one of the community moderators for Modular on the Discord server and Discourse forum.

Connect with Owen Hilyard

Seth Stadick

I'm a Bioinformatics Software Engineer passionate about building high-performance systems. For the past six years, I've developed in Rust across environments ranging from Raspberry Pi devices to full-scale HPC clusters in the cloud. I'm excited about Mojo's potential to reshape how we approach performance-critical computing.When I'm not coding, I’m usually spending time with my family — or trying to land new tricks on a skateboard (I just learned to Ollie)!

Connect with Seth Stadick

Brian Grenier

I am C++ developer working for a cardiac image processing company. I've been a Mojo standard library contributor since 2024. I also actively maintain two libraries, EmberJSON, a JSON library written in pure Mojo, and Kelvin, a type safe dimensional analysis library. I often hang out in the Mojo discord and forum!

Connect with Brian Grenier

Martin Vuyk

I'm a Mechatronics Engineer who pivoted into Software Development. I like tackling complex problems and building lasting solutions. I invest my time into things that I think will be impactful.

Connect with Martin Vuyk

Sawyer Bergeron

Compiler, PL, and systems/performance engineering enthusiast with an eyebrow-raising amount of VAX assembly experience. Collector (of hobbies). Enjoys bagels just a bit too much.

Connect with Sawyer Bergeron

Valentin Erokhin

Author of Lightbug, a Mojo HTTP Framework (https://github.com/Lightbug-HQ/lightbug_http)

Connect with Valentin Erokhin

Sora

Connect with Sora

Maxim Zaks

I tell computers how to waste electricity, hopefully in an efficient or at least useful way.

Connect with Maxim Zaks

Gabriel de Marmiesse

Former core dev of Keras, author of Python-on-whales, currently working at Kyutai to democratize AI through open science

Connect with Gabriel de Marmiesse

Max Brylski

Explorer of the computational universe

Connect with Max Brylski

Tilli Fe

Connect with Tilli Fe

Return to page

Leadership team

X/X

Chris Lattner

Distinguished Leader who founded and scaled critical infrastructure including LLVM, Clang, MLIR, Cloud TPUs and the Swift programming language. Chris built AI and core systems at multiple world leading technology companies including Apple, Google, SiFive and Tesla.

Connect with Chris Lattner

Tim Davis

Repeat Entrepreneur and Product Leader. Tim helped build, found and scale large parts of Google's AI infrastructure at Google Brain and Core Systems from APIs (TensorFlow), Compilers (XLA & MLIR) and runtimes for server (CPU/GPU/TPU) and TF Lite (Mobile/Micro/Web), Android ML & NNAPI, large model infrastructure & OSS for billions of users and devices. Loves running, building and scaling products to help people, and the world.

Connect with Tim Davis

Mostafa Hagog

Mostafa is a seasoned engineering leader in high-performance computing. During his tenure at NVIDIA, he served as Engineering Director and led teams to develop optimized deep learning libraries like cuDNN and CUTLASS, revolutionizing GPU-accelerated AI. At SiFive, as VP of Software, Mostafa assumed a leadership role guiding teams in the development of an MLIR/LLVM-based software stack for SiFive Intelligence & performance cores. His contributions also extend to optimizing Intel GPU hardware/software features, playing a pivotal role in developing the AVX1/2 SIMD ISA for Intel CPUs, and contributing to the GNU C Compiler. Mostafa holds a Master of Science in Electrical Engineering from the Technion, with a specialization in compiler optimizations. His unwavering passion for innovation continues to drive advancements in the field of high-performance computing.

Connect with Mostafa Hagog

Kalor Lewis

Kalor is Modular's VP, Finance and leads all our Finance operations. Prior to Modular, Kalor was a VP, Finance at Fivetran where he was the first finance hire in 2018 and built out the companies entire finance function. Before Fivetran, Kalor was part of Palantir Technologies, where he scaled their strategic finance function.

Connect with Kalor Lewis

Eric Johnson

Product leader who has built and scaled AI applications and infrastructure. Eric led the TensorFlow API, Compiler, and Runtime teams at Google Brain and Core Systems, including the founding of TFRT and the productionization of JAX. He holds an MBA from Wharton and Computer Science MS from Penn and loves soccer, fitness, and the great outdoors.

Connect with Eric Johnson

Mike Edwards

Mike has spent over 25 years working in the fields of IT, corporate operations, and software development - most recently at Apple. Mike volunteers his time serving as a Board member with the LLVM Foundation, focusing on finance and operations. Mike truly believes in the power of AI to help address some of the world’s greatest needs.

Connect with Mike Edwards

Return to page