Hamza Abbasi
Building AI systems for high-stakes domains where accuracy, traceability, and reliability matter.
About Me
Principal Technical Architect
Automotive AI (AAI) GmbH
Building AI systems for high-stakes domains where accuracy, traceability, and reliability matter.
Over the past few years, I've worked across autonomous driving, safety-critical validation, and generative AI products.
I've led the architecture and development of platforms used by automotive stakeholders to validate ADAS functions, generate and execute driving scenarios, and improve the way engineers interact with complex technical standards.
Most recently, I've been building GenAI products end-to-end, including RAG pipelines, fine-tuning (LoRA/QLoRA), evaluation workflows (online and offline), and retrieval optimization using metrics like Precision@K and Recall@K.
I enjoy working close to real users, turning messy requirements into scalable systems, and shipping products that people actually trust.
Technical Skills
A comprehensive toolkit for building production AI systems
Programming
GenAI & LLMs
AI/ML
Automotive Standards
Cloud & Infrastructure
Data & Backend
Featured Projects
Production AI systems for automotive compliance, legal tech, and autonomous driving
CORA
Compliance & Regulatory Assistant
A GenAI assistant built to help homologation engineers navigate compliance documents and homologate vehicles faster. Features grounded question-answering with strong citation and traceability requirements.
Key Highlights
Tech Stack
OmniLegis
GenAI Legal Rights Platform
A GenAI-powered platform helping German B2C users understand their legal rights. Features production RAG pipelines with fine-tuned models for domain-specific legal accuracy.
Key Highlights
Tech Stack
ReplicaR
Scenario-Based Validation Platform
End-to-end scenario generation and validation platform for automated driving systems. Enables scenario-based validation aligned with ISO 34502 and ISO 34503 standards.
Key Highlights
Tech Stack
GoldenAlgo
ADAS + Self-Driving Validation (Audi GmbH)
Comprehensive ADAS and self-driving validation platform developed in collaboration with Audi GmbH. Features computer vision pipelines, sensor fusion, and KPI benchmarking.
Key Highlights
Tech Stack
Interested in collaboration?
I build AI systems where accuracy, traceability, and reliability matter. Let's discuss how we can create production-ready solutions together.
Get in TouchProfessional Experience
Building production AI systems for automotive and legal domains
Principal Technical Architect
CORA (Compliance & Regulatory Assistant)
GenAI for Homologation Engineers
- Led the architecture and development of CORA, a GenAI assistant for navigating compliance documents
- Managed and structured a large corpus of regulations from EUR-Lex, ISO, UN Regulations, and KBA
- Designed LLM services for document understanding, grounded Q&A, and structured compliance guidance
- Built RAG pipelines with strong citation, traceability, and reliability requirements
- Developed evaluation-driven workflows to improve accuracy and reduce hallucinations
OmniLegis
GenAI Legal Rights Platform (Germany)
- Led end-to-end architecture of a GenAI platform helping German B2C users understand legal rights
- Built production RAG pipelines using LangChain and LangGraph with MongoDB vector indexes
- Developed fine-tuning workflows using LoRA and QLoRA for domain-specific legal accuracy
- Implemented online/offline evaluation pipelines using LangSmith with Precision@K, Recall@K metrics
- Deployed and scaled the system on GCP, launched on App Store and Google Play Store
Technologies & Tools
AI and Platform Developer
ReplicaR
Scenario-Based Validation + TÜV SÜD Certified Platform
- Contributed to design and end-to-end integration for scenario-based ADAS validation
- Defined software architecture and supported ADAS Operational Design Domain (ODD) engineering
- Developed critical workflows for generating and executing driving scenarios (ISO 34502, ISO 34503)
- Worked with OpenDRIVE (XODR), OpenSCENARIO (XOSC), and OSI sensor simulation standards
- Built CI/CD pipelines using Docker, GitHub, Jenkins, and Kubernetes
- Helped drive technical decisions that led to TÜV SÜD fit-for-purpose certification
GoldenAlgo (Audi GmbH)
ADAS + Self-Driving Validation Platform
- Function owner for multiple ADAS and self-driving validation features with Audi GmbH
- Led large-scale data workflows including dataset handling, labeling, and specifications
- Managed model retraining cycles and performance monitoring for production CV systems
- Developed KPI and benchmarking pipelines in C++ for validation and quality monitoring
- Integrated raw automotive data including CAN, ECU signals, LiDAR, and camera streams
- Built Python pipelines for 3D perception, sensor fusion, and visual odometry
Technologies & Tools
Let's Connect
Interested in collaboration or just want to discuss AI systems? I'd love to hear from you!
Get in Touch
LinkedIn Preferred
For professional inquiries and networking, connect with me on LinkedIn for the fastest response!
Connect on LinkedInOpen to Collaboration
Consulting
AI strategy and implementation guidance
Projects
GenAI and automotive AI solutions
Speaking
Tech talks and workshops on AI
Research
Collaborative research opportunities
Send a Message
Response Time
I typically respond within 24-48 hours. For urgent matters, LinkedIn is your best bet!