Michael Pfaffenberger Leaves Walmart for Pydantic

Le brief IA que les pros lisent chaque soir
Les 7 actus IA du jour, décryptées en 5 min. Gratuit.
Inclus dès l'inscription : notre sélection des meilleurs guides & comparatifs IA.
Choisis ton rythme
Gratuit · Pas de spam · Désabonnement en 1 clic
Michael Pfaffenberger Leaves Walmart for Pydantic
Michael Pfaffenberger, the mastermind behind Walmart's coding assistant tool, Code Puppy, has decided to leave the company to join the AI startup Pydantic. This decision was communicated in an internal memo, which revealed that Code Puppy had reached an impressive user base of nearly 75,000 within Walmart, impacting all levels of the organization.
The Impact of Code Puppy at Walmart
Code Puppy was designed to enable Walmart employees, whether they are engineers or regular store associates, to create their own software tools. This tool is based on the Pydantic AI library and was developed by Pfaffenberger while he served as a distinguished engineer within Walmart's Global Tech group. With this AI-powered coding assistant, developers can write, edit, test, and manage software using natural language instructions.
A Thoughtful Departure
In his memo, Pfaffenberger expressed his satisfaction in having created software that improved the lives of many employees, both professionally and personally. He acknowledged that leaving Walmart was not an easy decision, but it was time for him to take on new challenges.
Walmart's Reaction
Walmart confirmed Pfaffenberger's departure and wished him the best for his future endeavors. The company is currently in a quiet period ahead of its financial results announcement scheduled for August 20.
The Adoption of Code Puppy
Suresh Kumar, Walmart's Chief Technology Officer, stated in June that Code Puppy was being used in various areas, such as merchandising and supply chain. He noted that the tool had democratized access to software development within the company, with the number of non-engineer users now comparable to that of engineers, including hourly associates.
A Concrete Example of Use
A regional load manager used Code Puppy to create a tool capable of analyzing hundreds of available truck loads in the region, aiming to minimize empty returns for drivers after a delivery.
A Strategy for Technological Independence
Code Puppy is part of Walmart's efforts to avoid excessive reliance on a limited number of AI suppliers. Unlike tools like Claude Code from Anthropic or Codex from OpenAI, Code Puppy is not tied to a single model or AI provider. It can work with various models from different suppliers, allowing developers to switch models, compare results, or use multiple models simultaneously. This flexibility also helps distribute workloads among suppliers, avoiding usage limits and controlling costs.
Well-Deserved Recognition
At last month's annual shareholder meeting, Pfaffenberger and his colleague John Choi were recognized by CEO John Furner for their innovative project.
The Future at Pydantic
David Montague, Chief Technology Officer of Pydantic, stated that Pfaffenberger brings valuable experience gained from a large company, building AI tools capable of coordinating multiple agents. This skill is essential as large enterprises seek to adopt AI in a more integrated manner.
Pfaffenberger's Memo to His Team
In his farewell message, Pfaffenberger recalled that a year ago, Code Puppy had barely 1,000 users. Today, that number is approaching 75,000. He expressed his gratitude to the users who have been his source of motivation and emphasized that the decision to leave was not easy.
Pfaffenberger also thanked Brian Knapp for his support in building a strong team that will continue to advance the project in his absence. He named several key members of this team, including Ryan Eccles, John Choi, Kaveri Bakshi, Wesley Blakemore, Andrew Tilson, and TJ Webb.
The Impact of Open Source
Pfaffenberger highlighted the importance of making Code Puppy open source, which has allowed it to be downloaded over 496,000 times, including 71,000 times last month. This openness has attracted contributions from developers worldwide, including students from prestigious universities like Stanford.
In September and October of the previous year, the open-source version of Code Puppy was widely used on Cerebras' Discord during the launch of their inference product. TJ Webb, who recently joined the team, was among those active users.
The open-source version of Code Puppy is available at github.com/mpfaffenberger/code_puppy. Pfaffenberger encourages users to try it at home, although it is not compatible with Walmart's internal systems.
Pfaffenberger concluded his message by expressing gratitude to everyone who contributed to the success of Code Puppy, sincerely thanking his colleagues and users for their unwavering support.
Brief IA — L'actualité IA en français
L'essentiel de l'actualité de l'intelligence artificielle, décrypté et expliqué chaque jour.