Home
WorkBlogTalksPapersAbout
Nikita Kozodoi

Blog

29Posts
Nikita Kozodoi

Shaping the Future of Embedded Automotive Development with AUMOVIO and AWS

  1. Home
  2. Blog
  3. Shaping the Future of Embedded Automotive Development with AUMOVIO and AWS
Nikita KozodoiSeptember 20254 min read

Catching AUTOSAR compliance problems while the code is written

GenAIAutomotive
Read the full post on AWS Blog

Automotive software has a documentation problem before it has an AI problem. Developers document, review and test every code change against standards like AUTOSAR, and even a small update can touch hundreds of files. They usually find out whether they got it right at the very end, when the compliance tools run and send them back to fix things.

AUMOVIO, the company Continental's automotive technology business became, builds a development platform on AWS called vSDV Composer that replaces physical test hardware with virtual ECUs. We worked with them on the engineering assistant inside it. It runs on Amazon Bedrock, and AUMOVIO trained it on their own knowledge base, covering more than twenty years of embedded source code and process documentation.

What makes it useful is how narrow it is. A general coding assistant has no opinion about AUTOSAR, while this one flags likely compliance problems in the IDE, while a developer is still writing the code, instead of at the end of the pipeline. Moving that feedback earlier is the whole point, since the cost of a compliance error is mostly the round trip rather than the fix. AUMOVIO reports the platform accelerating development by more than 50%.

Where this is heading is agents coordinating across organizations rather than within one tool, so an ECU project could pull in validation agents from its suppliers over protocols like A2A. That is harder than it first sounds, because the interesting part isn't the message format. It's working out whose answer to trust.

Nikita Kozodoi

Written by Nikita Kozodoi

Senior Scientist on the AWS FDE team. I publish applied AI research and embed with engineering teams to build custom AI systems, from scoping and experimentation to fine-tuning and production deployment.

See my workGet in touch

Related content

Blog

AUMOVIO Boosts Software Development with an Agentic Coding Assistant Powered by Amazon Bedrock

2026
Blog

AUMOVIO Improves Quality of Automotive Software at Scale Using Multi-Agent AI on Amazon Bedrock

2026
Blog

Accelerate Medical Content Review with Amazon Bedrock AgentCore

2026

Pages

  • Work
  • Blog
  • Talks
  • Papers
  • About

Social

  • LinkedIn
  • GitHub
  • Google Scholar
  • X / Twitter
  • Instagram

Contact

  • n.kozodoi@icloud.com
  • Buy me a coffee
  • Download CV
  • RSS feed
  • Berlin, Germany

© 2026 Nikita Kozodoi. All opinions are my own.