Hey, I'm Noah.
I'm an AI Systems Engineer and ML Engineer studying for an M.Sc. in Artificial Intelligence at UC Berkeley & NTNU. I build production-grade AI systems, multi-agent pipelines, and scalable backend infrastructure.
Currently
- Finishing M.Sc. coursework at UC Berkeley — NLP, recommender systems, big data, neurotechnology
- Research on in-context learning robustness, carbon-aware recommendations, and healthcare accessibility modeling
- Latest vLLM benchmark finding: increasing concurrency from 1 to 8 mainly raised queueing latency, and doubling max_tokens from 128 to 256 roughly doubled latency while tokens/sec stayed similar
- Stepping into Head of Technology at ReLU NTNU
Experience
AI/Software Engineering Intern
Spacial AI · Palo Alto, CA
IT Consultant Intern
Bouvet · Oslo, Norway
Learning Assistant
NTNU · Trondheim, Norway
Education
UC Berkeley
2025 – PresentExchange Year · GPA 3.65/4.0
NTNU
2022 – PresentM.Sc. Artificial Intelligence · B.Sc. Informatics (GPA 4.2/5.0)
Awards
1st Place — DigitalOcean Hackathon, San Francisco
2nd Place — Future of Labor Hackathon, San Francisco
Leadership
Head of Technology — ReLU NTNU
Norway's leading AI student organization. Driving tech choices for projects and keeping the organization at the forefront of AI/ML.
CEO — Datakameratene FK
Led the club to a league championship. Finance, sports administration, and team operations.
2023 – 2025Projects
view all →LLM Inference Benchmark
Open-source benchmark harness for comparing LLM inference performance. Latest vLLM result: on local single-worker Qwen2-0.5B-Instruct runs, higher concurrency mainly increased queueing latency while tokens/sec remained similar.
Carbon-Aware Recommender System
Studying the Pareto frontier between user engagement and carbon footprint in recommendation systems.
Skills
AI/ML: LLMs, RAG, agentic systems, NLP, recommender systems, RL (MuZero, MCTS)
Backend & Systems: Python, distributed systems, APIs, caching, observability
Cloud: AWS (Lambda, API Gateway, S3, Bedrock), Docker, GCP
Languages: Python, TypeScript, Java, SQL