Shulin Gao

AI & Data Engineer · LLM Systems · Cloud · Berlin

---

About

• 🏭 10 years production AI/ML at Eurailscout — deep learning systems for national-scale railway inspection across optical, ultrasound & eddy-current sensors
• 🤖 Now building agentic LLM systems and cloud-native data pipelines that bridge foundation models with real-world sensor and business data
• 🎓 M.Sc. Computational Science & Engineering, TU Munich (Grade 1.8) · Published researcher (ECNDT 2018 first author · WCNDT 2016)

Skills: Python · C++ · PyTorch · CUDA · TensorRT · LLM · RAG · multi-agent systems · FastAPI · AWS · Terraform · Docker · PostgreSQL · MongoDB · RabbitMQ · OpenCV · MLflow · data pipelines · computer vision · HPC

💼  Open to AI Engineer and Data Engineer roles in Berlin

---

Tech Stack

LLM / Agents

Cloud

Data

ML / Vision

Languages

---

GitHub Activity

---

Featured Projects

Project	Description	Stack
iHunter	Agentic LLM platform for B2B lead discovery — multi-agent orchestration, automated data enrichment & quality assessment	Python · Groq · LLaMA 3 · FastAPI · aiosqlite
landlord_system	Property management for German landlords — rent tracking, utility billing, and PDF generation	Python · PostgreSQL · FastAPI
AWS Pedestrian Detection	Event-driven ML pipeline: S3 → SNS → Lambda → RDS · Terraform IaC	AWS · Python · HCL
MonoSLAMAutoCalibration	M.Sc. thesis — monocular SLAM for automatic camera self-calibration of surgical endoscopes (Gaussian Sum Filter)	C++ · OpenCV · Eigen
kleinanzeigen-scraper	Full-stack scraper with browser UI, price & location analytics	Python · FastAPI · PostgreSQL

---

Publications

• S. Gao et al. — Multi-modal sensor fusion for automated railway defect classification · ECNDT 2018 · First author
• S. Gao et al. — Cross-sensor data fusion for rail surface analysis · WCNDT 2016

---

Languages

🇬🇧 English — C2  ·  Cambridge BEC Higher  |  🇩🇪 German — C1  |  🇨🇳 Chinese (Mandarin) — Native

---

Connect