• Google Scholar
• LinkedIn
• YouTube
• Document

Ali Nouri
AI Researcher @ Chalmers University of Technology
Arguing Safety of Autonomous Vehicle @ Volvo Cars

---

Research Goal:
Accelerating DevSafeOps for Autonomous Driving Software.
Collaboration between Volvo Cars, Zenseact, and Chalmers.

---

Recent Event: Public Defence
Topic: Accelerating the Design Phase -
Towards DevSafeOps for Autonomous Driving Software
Opponent: Prof. Philip Koopman, Carnegie Mellon University, US

Research Abstract

Background:

The safety of Autonomous Driving (AD) remains a barrier to its widespread adoption, as evidenced by recent incidents. Factors such as the complex environment, evolving technologies, and shifting regulatory and customer requirements necessitate continuous monitoring and improvement of AD software (Fig. 1). This is a process that may favor software and system engineering supported by DevOps. The iterative DevOps process is crucial, serving two purposes: satisfying customer demands through continuous improvement of the function and providing a framework for timely responses to unknown bugs or incidents.

Fig. 1: Autonomous Driving DevOps process (Publication A)

However, any update to the software must follow rigorous safety processes prescribed by standards, regulations, or the state of the art in industry. Incorporating these safety activities into the DevOps forms an iterative process called DevSafeOps. These necessary activities, although vital for safety assurance, inherently lead to a compromise in rapidity.

Research Goal:

In our research, we initially identify the challenges of rapid DevSafeOps in AD development, and then explore existing solutions. Subsequently, we propose multiple approaches for the acceleration of safety analysis, requirements engineering, code generation, and synthetic data generation in DevSafeOps cycles.

Methods:
To address each research objective, diverse research methods are utilized. Interview studies and a systematic literature review are conducted to identify the challenges, research gaps, and existing approaches. Then, design science, interview study, case study, and experimentation are employed.Results:
Initially, the challenges and research gaps related to each essential activity for the safety of automated driving are identified (Papers A and B), together with the proposed solutions presented in the literature (Paper B).

Then, two approaches are proposed to accelerate safety-concept design (i.e., analysis and requirements engineering) as an initial step in DevSafeOps. We adapt System Theoretic Process Analysis (STPA) to enable distributed development within automotive system engineering (Paper C).

Fig. 2: Distributed development (Publication C)

As an alternative approach, a large-language-model (LLM)-based multi-agent Hazard Analysis and Risk Assessment (HARA) prototype is developed and evaluated to enable automation (Papers D and E). LLM-based HARA (Fig. 3), utilizing a pipeline of subtasks, each managed through a specific prompt. The item definition is imported (top-left), and the HARA results are exported (bottom-right). In the second row of the HARA table, the relationship of each column to the prompts is summarized.

Fig. 3. LLM-based Multi-Agent HARA

Then, the rule-based software-implementation phase is accelerated through LLM-based code generation conducted through a conversation in a simulation environment (Papers F and G). Initially, The LLM-generated code is evaluated automatically in a simulation model against multiple critical traffic scenarios, and an assessment report is provided as feedback to the LLM for modification or bug fixing. We report about the experimental results of the prototype employing Codellama:34b, DeepSeek (r1:32b and Coder:33b), CodeGemma:7b, Mistral:7b, and GPT4 for Adaptive Cruise Control (ACC) and Unsupervised Collision Avoidance by Evasive Manoeuvre (CAEM).

Current research:

Conclusions:
My thesis first identifies multiple challenges in achieving rapid DevSafeOps in AD development and then proposes several approaches for addressing these challenges across different phases of the DevSafeOps cycle. To accelerate the design phase, we introduce an adaptation of STPA for multiparty distributed development and employ multi-agent LLMs as a parallel approach for HARA. We further examine how LLMs and VLMs can support safety concept design, code generation, and monitoring activities with reduced engineer involvement, while defining the necessary safeguarding measures. Finally, we investigate 3DGS as an effective and rapid DataOps technique within DevSafeOps, enabling improved data generation and augmentation for ML-based software development.

Recent Award

CHALLENGE AWARD IN RE '24: "Expanding the Frontiers of RE"

Recent Publications

• On Simulation-Guided LLM-based Code Generation for Safe Autonomous Driving Software

A. Nouri, J. Andersson, K. D. J. Hornig, Z. Fei, E. Knabe, H. Sivencrona, B. Cabrero-Daniel, C. Berger
EASE 2025, 29th International Conference on Evaluation and Assessment in Software Engineering.

• Large Language Models in Code Co-generation for Safe Autonomous Vehicles

A. Nouri, B. Cabrero-Daniel, Z. Fei, K. Ronanki, H. Sivencrona, C. Berger
4th International Conference on Computer Safety, Reliability, and Security. SAFECOMP 2025. Lecture Notes in Computer Science, vol 15954. Springer, Cham.

• The DevSafeOps Dilemma: A Systematic Literature Review on Rapidity in Safe Autonomous Driving Development and Operation

A. Nouri, B. Cabrero-Daniel, F. Torner, C. Berger
Journal of Systems and Software, Volume 230, 2025, 112555, ISSN 0164-1212.

• Engineering Safety Requirements for Autonomous Driving with Large Language Models

A. Nouri, B. Cabrero-Daniel, F. Torner, H. Sivencrona, C. Berger
2024 IEEE 32nd International Requirements Engineering Conference (RE), Reykjavik, Iceland, 2024, pp. 218-228.

• Welcome Your New AI Teammate: On Safety Analysis by Leashing Large Language Models

A. Nouri, B. Cabrero-Daniel, F. Torner, C. Berger
IEEE/ACM 3rd International Conference on AI Engineering - Software Engineering for AI (CAIN '24). Association for Computing Machinery, New York, NY, USA, 172–177.

• An Industrial Experience Report about Challenges from Continuous Monitoring, Improvement, and Deployment for Autonomous Driving Features

A. Nouri, C. Berger, F. Torner
2022 48th Euromicro Conference on Software Engineering and Advanced Applications (SEAA), Gran Canaria, Spain, 2022, pp. 358-365.

• On STPA for Distributed Development of Safe Autonomous Driving: An Interview Study

A. Nouri, C. Berger, F. Torner
2023 49th Euromicro Conference on Software Engineering and Advanced Applications (SEAA), Durres, Albania, 2023, pp. 5-12.

• An LLM and Embeddings-Based Multi-agentic System for Knowledge Graph Construction and Verification

M. R. Martínez Rodríguez, A. Nouri, Z. Fei & M. M. Hedblom
PRIMA 2025: Principles and Practice of Multi-Agent Systems. PRIMA 2025. Lecture Notes in Computer Science(), vol 16366. Springer, Cham.

Patents (publicly available)

Upcoming Presentations

• VECS 2026 - Accelerating DevSafeOps for Autonomous Driving through Generative AI: Silver Bullet or Just Buzz?

Selected Previous Presentations

• MIT 2022 STAMP Workshop - A feasibility study on the application of STPA method in AD
• Auto AI USA (San Francisco) - Accelerating AV DevSafeOps - Large Language Models as a potential approach
• Autonomous Vehicle Technology Expo 2024 - Leashing AI: On Challenges and Solutions in Autonomous Vehicle Safety Assurance
• CAIN 2024 - Welcome Your New AI Teammate: On Safety Analysis by Leashing Large Language Models
• Keynote @ SRToITS 2025 is in conjunction with SAFECOMP 2025 - DevSafeOps for Autonomous Driving: Generative AI

Guest Lecture

Supervision Activities

Experiences

Education