Herogen signals a shift to autonomous AI-driven development, enabling Delivery Hero engineers to scale output, reduce manual coding, and focus on high-impact innovation
Delivery Hero has introduced a transformative leap in software engineering with the unveiling of Herogen, an autonomous AI-powered software delivery agent that is already reshaping how code is written, reviewed, and deployed at scale. Positioned as a virtual engineering force multiplier, Herogen currently delivers output equivalent to a team of approximately 130 senior engineers—and its capabilities continue to expand rapidly as adoption grows across the organization.
The announcement signals a broader shift in how large technology-driven enterprises approach software development. Rather than treating artificial intelligence as a productivity aid, Delivery Hero is operationalizing AI as an autonomous contributor embedded directly within engineering workflows.
A New Paradigm in Software Development
Herogen was first rolled out internally in February and has since been gradually introduced across Delivery Hero’s global product and engineering teams. Despite being only 18% deployed across the organization, the system is already responsible for approximately 9% of all code change requests—a remarkable показатель of early impact.
Built in-house using advanced Large Language Models (LLMs), Herogen is designed to function not as a passive assistant but as an active, autonomous agent capable of executing complex software delivery tasks end-to-end. Engineers interact with the system using natural language, describing the intent of a task rather than writing code line by line.
From there, Herogen independently:
- Generates production-ready code
- Conducts testing and validation
- Iterates on improvements
- Submits finalized pull requests for review
This workflow fundamentally alters the role of engineers, shifting their focus from implementation details to higher-level system design and innovation.
The “Council of Agents” Architecture
A defining feature of Herogen is its multi-model validation system, referred to internally as a “council of agents.” Rather than relying on a single AI model, Delivery Hero has architected a layered review process in which multiple LLMs—sourced from different providers—evaluate generated code from diverse perspectives.
This architecture introduces redundancy and reduces systemic bias or blind spots that may exist within any individual model. Each agent in the council assesses factors such as:
- Code correctness
- Security vulnerabilities
- Performance efficiency
- Architectural consistency
Only after passing through this multi-agent review does the code reach a human engineer for final validation. This hybrid approach ensures that autonomy does not come at the expense of quality or governance.
The results have been notable. Herogen currently achieves an 85% success rate, measured by the proportion of generated pull requests that are successfully merged versus those that are rejected. Furthermore, the majority of tasks require minimal human interaction—often zero or just a single review cycle—demonstrating a high level of operational maturity.
Scaling Output Without Expanding Headcount
One of the most striking outcomes of Herogen’s deployment is its impact on engineering productivity. By autonomously merging over 100 code change requests per day, the system has effectively eliminated a substantial portion of repetitive development work.
Delivery Hero estimates that Herogen is already saving approximately 250,000 hours of manual coding annually. This reclaimed capacity allows engineers to redirect their efforts toward more strategic initiatives, including system architecture, feature innovation, and long-term platform evolution.
According to Chief Technology Officer Benjamin Mann, this shift is not about replacing engineers but amplifying their impact. By offloading routine tasks to Herogen, the company enables its existing talent to operate at a higher level of abstraction, tackling more complex and value-generating challenges.
This aligns with Delivery Hero’s broader ambition to evolve into a global “Everyday App,” a platform that integrates food delivery, groceries, and other daily services into a unified digital experience. Achieving this vision requires rapid iteration and continuous innovation—both of which are accelerated by Herogen’s capabilities.
From Code to Intent: Redefining the Developer Interface
Herogen represents a fundamental shift in the interface between humans and machines. Traditionally, software development has required engineers to translate ideas into precise programming languages—a process that is both time-consuming and cognitively demanding.
With Herogen, this paradigm is inverted. Engineers express their intent in natural language, and the system handles the translation into executable code. This abstraction layer reduces friction in the development process and enables faster experimentation.
Rodrigue Schäfer, Vice President of Platform at Delivery Hero, emphasizes that this transition allows engineers to focus on creative problem-solving rather than syntactic correctness. By removing the need to “think in code,” developers can operate more intuitively, exploring ideas and iterating rapidly.
This approach also lowers barriers to collaboration, as non-specialist stakeholders can more easily articulate requirements and contribute to the development process.
Engineering Evolution, Not Replacement
Despite its advanced capabilities, Herogen is not positioned as a replacement for human engineers. Instead, it represents a shift in how engineering roles are defined and executed.
Traditional automation often aims to eliminate repetitive tasks, but Herogen goes further by redefining the nature of those tasks altogether. Engineers are no longer primarily responsible for writing code; they are becoming orchestrators of systems, designers of workflows, and stewards of architectural integrity.
This evolution mirrors broader trends in the software industry, where senior engineers increasingly focus on system-level thinking rather than individual components. Herogen accelerates this transition by enabling even mid-level developers to operate at a higher level of abstraction.
The result is a workforce that is more strategic, more creative, and more aligned with business outcomes.
Enterprise-Scale Impact and Future Expansion
While Herogen’s current deployment covers less than one-fifth of Delivery Hero’s engineering teams, its impact is already significant. The company plans to expand its adoption aggressively, with a target of handling 20% of all code change requests by the end of the year.
As adoption increases, the system’s contribution to overall engineering output is expected to grow exponentially. This scalability is a key advantage of AI-driven systems, which can be replicated and expanded without the constraints of traditional hiring and onboarding processes.
Moreover, the data generated by Herogen’s operations—such as code patterns, performance metrics, and failure cases—can be used to further refine and improve the system, creating a feedback loop that enhances its capabilities over time.
Implications for the Broader Industry
Delivery Hero’s deployment of Herogen offers a glimpse into the future of software engineering. As AI systems become more autonomous and reliable, the distinction between human and machine contributions will continue to blur.
Key implications include:
- Increased productivity: Organizations can achieve higher output without proportional increases in headcount
- Faster innovation cycles: Reduced development time enables rapid experimentation and iteration
- Shift in skill requirements: Engineers will need to focus more on system design, critical thinking, and domain expertise
- Enhanced collaboration: Natural language interfaces make development more accessible to a broader range of stakeholders
At the same time, challenges remain. Ensuring security, maintaining accountability, and managing the complexity of AI-driven systems will require robust governance frameworks and continuous oversight.
With the launch of Herogen, Delivery Hero is pioneering a new model of software development—one in which autonomous AI agents operate alongside human engineers to deliver unprecedented levels of productivity and innovation.
By combining advanced LLMs with a sophisticated multi-agent review system, Herogen achieves a balance between autonomy and reliability that has historically been difficult to attain. Its ability to generate, test, and deploy code with minimal human intervention represents a significant step forward in the evolution of AI-driven engineering.
As the platform continues to scale, it is likely to play a central role in Delivery Hero’s transformation into a global Everyday App, while also setting a precedent for how enterprises across industries can harness AI to redefine their development processes.
In this emerging landscape, the question is no longer whether AI will participate in software engineering—but how deeply it will be integrated into the fabric of development itself.
Source link: https://www.businesswire.com
