📍 Karlsruhe, Germany · 🇵🇰 → 🇩🇪
name: Hamza Aamer
role: Senior AI Engineer @ Aiden Risk # AI-native insurance brokerage
research: HiWi @ KIT — CES/ITEC & TECO # LLM inference, edge ML
studying: M.Sc. Computer Science (Human-centred Machine Intelligence) @ KIT
based_in: Karlsruhe, Germany
focus: [multi-agent systems, synthetic data, LLM inference, computer vision]
philosophy: "Production over demos."I build AI systems that hold up once real users, real data and real compliance rules arrive — agent orchestration, rule-based data synthesis for regulated domains, and squeezing models onto hardware that wasn't designed for them.
|
🛡️ Aiden Risk Building the AI layer of an AI-native insurance brokerage: agents that read, reason over and act on insurance documents. |
🔬 KIT Research Research on efficient LLM inference and ML on edge / embedded devices across two chairs. |
🎓 M.Sc. at KIT Computer Science with a focus on Human-centred Machine Intelligence. |
+ 2026 ── Senior AI Engineer ··························· Aiden Risk
+ 2026 ── Research Assistant (HiWi) ···················· KIT — CES/ITEC & TECO
! 2025 ── AI Engineer & Product Lead ··················· Adept Tech Solutions
! 2024 ── AI Engineer ·································· Antematter
- 2023 ── ML Engineer ·································· DataInsight🛡️ Senior AI Engineer — Aiden Risk · 2026 – Present
- Joining the core AI team at an AI-native insurance brokerage, working on agentic workflows for broking operations.
🔬 Research Assistant (HiWi) — Karlsruhe Institute of Technology · Apr 2026 – Present
- Chair for Embedded Systems (CES / ITEC) and TECO (Pervasive Computing)
- Research on LLM inference efficiency and edge ML deployment
⚡ AI Engineer & Product Lead — Adept Tech Solutions · Feb 2025 – Sep 2026
- Led the rule-based synthetic data track of Ascend, a data synthesis platform built for a large US insurance carrier's claims systems
- Designed the DNA Signature Technology, capturing the statistical fingerprint of production datasets at 99.5% statistical fidelity
- Built PySpark synthesis pipelines processing 10M+ rows
- Improved geospatial analysis accuracy by 40% through distributed computing optimisation
- Co-authored the engagement's target-state business problems & outcomes document
🤖 AI Engineer — Antematter · Jul 2024 – Feb 2025
- Built Ant-AI, a genetic multi-agent framework that cut prompt ambiguity by 70%
- Shipped multi-agent systems with CrewAI, LangChain, ELIZA and Phidata, including supervisory agents
- Automated outreach pipelines for a 10× increase in company outreach
👁️ ML Engineer — DataInsight · Jun 2023 – Sep 2023
- Contributed to a $90K project using YOLO + image processing for text recognition
- Sped up dataset acquisition 2.5× with TensorFlow and OpenCV
|
Synthetic data platform · Adept Tech Solutions Rule-based synthesis for regulated insurance data, driven by a statistical "DNA signature" of the source.
99.5% fidelity · 10M+ rows |
Genetic prompt optimiser · Nov 2024 Multi-agent framework where prompts evolve through genetic selection until they stop being ambiguous.
70% ambiguity reduction |
|
3D neural reconstruction for VR · Final Year Project Real-time environment reconstruction with 3D Gaussian Splatting, rendered on an HTC Vive.
90 FPS · −60% memory · PSNR 28.5 |
Aerial roof analysis · May 2024 Roof measurement from satellite imagery fused with LiDAR point clouds, served through a GIS pipeline.
95% classification accuracy |
|
Image-to-image translation · Jul 2024 CycleGAN and conditional GAN experiments for paired and unpaired domain translation.
|
Systems & mobile ML · 2023 A multi-core neural network runtime written in C, and a shopping app with an on-device vision model.
|
-0D1117?style=flat-square){kind=icon}
| Domain | What I reach for |
|---|---|
| Agents | CrewAI · LangChain · Phidata · ELIZA · supervisory / multi-agent orchestration |
| Data | PySpark · rule-based synthesis · statistical modeling · PostgreSQL · Redis |
| ML / DL | PyTorch · TensorFlow · scikit-learn · CUDA · GANs · Gaussian Splatting |
| Vision & Docs | YOLOv8 · OpenCV · Open3D · Donut · DocParser |
| Efficiency | LLM inference optimisation · edge / embedded ML |
| Ship it | FastAPI · Flask · Streamlit · Docker · Linux |
| Degree | Institution | Years | |
|---|---|---|---|
| 🇩🇪 | M.Sc. Computer Science — Human-centred Machine Intelligence | Karlsruhe Institute of Technology (KIT) | 2025 – 2027 |
| 🇵🇰 | B.S. Computer Science | FAST NUCES | 2021 – 2025 |
Open to research collaborations on LLM inference & edge ML, and to conversations about agentic systems in regulated industries.
"Building AI systems that thrive in production, not just demos."