Happy Holidays & Merry Christmas New office finally updated with new RGB sign 🎄
Code Metal
Software Development
Boston, MA 2,465 followers
AI-Powered development at the Edge.
About us
“The future is already here. It's just not evenly distributed yet”
- Website
-
https://codemetal.ai
External link for Code Metal
- Industry
- Software Development
- Company size
- 11-50 employees
- Headquarters
- Boston, MA
- Type
- Privately Held
- Founded
- 2023
- Specialties
- Artifical Intelligence, Large Language Models (LLM), Edge Processing, Embedded Systems, Code Generation, Compilers, and AI Agents
Locations
-
Primary
Boston, MA, US
Employees at Code Metal
Updates
-
Code Metal reposted this
Code Metal Appoints Ayon Basumallik as SVP of Engineering and Product Learn more & get our take 👇 https://lnkd.in/gHgp4N8e “Ayon’s experience spans fundamental physics, firmware, and production-grade machine learning, making him ideally suited to Code Metal’s mission to bridge software and hardware for seamless deployment at the edge and help manage the influx of customers.” — Peter Morales, CEO & Founder at Code Metal “As edge use cases become more prevalent, the ability to translate common programming languages into code that is optimized for the edge will be increasingly important. Code Metal is at the forefront of tackling this challenge.” — Ayon Basumallik, Senior Vice President of Engineering and Product at Code Metal #SoTNews
-
Thanks Andrew Côté for Organizing the (Defense Tech Hackathon) https://lnkd.in/ghB6dyeB on the USS Hornet Air and Space Museum. We're having a blast supporting teams hacking on IoT devices and the Boston Dynamics SPOT (thanks for bringing spot Vitaly Bulatov!). Niranjan Hasabnis and Octavian Udrea did a little internal hackathon using some of our infrastructure to make it easy to program microcontrollers directly from python and transpile that to orders of magnitude faster C : https://lnkd.in/gXR-7t27 Looking forward to see what teams build!
-
Accelerate and deploy your #AI applications and #SignalProcessing chains on AMD FPGAs directly from #Python and #MATLAB. Connect with us: https://www.codemetal.ai/
-
Code Metal reposted this
It was a pleasure being on this panel last week. We had a deep discussion about the regulatory oversight of AI and positive impacts of AI use. My conclusion - like anything else -, AI impact will likely come down to the values and intentions of the people using it. Thank you to Joao-Pierre Ruth for inviting me and to Kjell Carlsson, Ph.D. and Sohrob Kazerounian for the discussion.
As Regulators Escalate Oversight, Who Are the AI Heroes? https://lnkd.in/e95BkEGh
As Regulators Escalate Oversight, Who Are the AI Heroes?
informationweek.com
-
Code Metal reposted this
Grateful for the industry collaboration with Intel Labs and Code Metal, contributing to award-winning work like MonoCoder at #IEEE #HPEC. Contributors: Tal Kadosh, Niranjan Hasabnis, Vy Vo, Nadav Schneider, Neva Krien, Mihai Capotă, Abdul Wasay, Guy Tamir, Ted Willke, Nesreen K. Ahmed, Yuval Pinter and Tim Mattson
At this year’s IEEE High Performance Extreme Computing Virtual Conference, we received the Outstanding Paper Award for MONOCODER, an HPC-specific language model. MONOCODER understands HPC code better than state-of-the-art LLMs to deliver greater performance across both non-HPC and HPC codes. Read the blog to learn more about the award-winning paper and the four other Intel papers submitted and accepted during the event. https://intel.ly/3N0dWYQ #Research #LLMs #HighPerformanceComputing
-
To celebrate SoCal Deep Tech week we're releasing the second in our series of research blog posts highlighting our team's work with University of California, Berkeley. LLMLift automatically translates programs written in languages like Java, C, and C++ into newer target languages, enabling them to run on the latest hardware. To ensure accuracy, LLMLift leverages a formal verification method (covered in our first blog post) to check the LLM outputs, catching any potential errors. Essentially, LLMLift is the first neuro-symbolic system designed for automatic code migration. https://lnkd.in/gcepGC5u
Combining AI with formal verification for efficient migration of legac
codemetal.ai
-
At CodeMetal, we're excited to share the first in a series of research blog posts! In this installment, we explore how our team's research, Tenspiler, automates the migration of legacy C++ and Python code to AI-specific DSLs like TensorFlow, PyTorch, and even more exotic DSLs like Apple MLX and Gemmini. This breakthrough not only simplifies the process but also delivers significant performance boosts on cutting-edge AI hardware. Stay tuned for follow-up articles on this and other topics. https://lnkd.in/gpZJeeY8
-
Very excited by the activity going on at Defense Advanced Research Projects Agency (DARPA) recently. The first compiler program in over 10 years kicked off today with Howard Shrobe and the MOCHA BAA. And Dan Wallach’s TRACTOR program is looking at new ways of rapidly moving the large legacy of C and C++ in our most important systems to Rust. Check out our own Peter Morales discussing his thoughts on the program https://lnkd.in/eMZFMYGb
DARPA suggests turning legacy C code automatically into Rust
theregister.com