# Martin Gottweis

**AI Agents & LLM Systems · Technical Lead**

New York, NY<br>
gottwik@gmail.com<br>
(+1) 650 250 7816<br>
[linkedin.com/in/martingottweis](https://www.linkedin.com/in/martingottweis/)


## Experience

### Software Engineer, Meta
*June 2021 - Present · New York, NY*

**AI Personalization**

- Led technical direction for AI personalization systems enabling privacy-aware, context-sensitive AI agent experiences across Meta products.
- Built graph-RAG infrastructure powering messaging search for ~500M users.
- Engineered static-knowledge context injection for WhatsApp AI, serving ~300K queries/day.
- Designed persistent-memory algorithms enabling AI agents to retain, retrieve, and act on user signals over time.
- Established architecture and implementation patterns adopted by the broader team.

**Integrity**

- Technical lead for two of Meta's core Integrity systems, including one operating at 8B+ QPS.
- Led changes to ultra-high-scale enforcement infrastructure where failures could affect harmful-content enforcement or platform availability across major Meta products.
- Delivered production changes at extreme scale despite limited access to standard testing, rollout, and gating infrastructure.

### Senior Software Developer, Booking.com
*April 2017 - June 2021 · Amsterdam, Netherlands*

- Drove 0-to-1 product development in Booking.com's new-ventures division, shipping prototypes into production for millions of travelers a day.
- Built targeted-discounting and real-time intent-capture systems used to personalize offers and improve booking conversion.
- Technical lead for Booking.com's hotel page, the core revenue funnel for a global travel platform, orchestrating 10K+ QPS across pricing, availability, personalization, merchandising, and accommodation-data systems to generate customized hotel pages at scale.

### Full-stack Software Developer, KISKA
*December 2015 - April 2017 · Salzburg, Austria*

- Built rapid-prototyping platforms and conceptual software, such as motorcycle dashboards, at the award-winning design studio behind the KTM motorcycle brand.

### Software Engineer, ASC Applied Software Consultants
*June 2007 - December 2015*

- Designed school-scheduling software for the Swiss market, solving complex real-world constraints such as teachers shared across multiple schools and travel times between campuses.
- Built internal tools and embedded systems for a market leader in school scheduling software.


## Skills

**Core competencies:** AI agents, LLMs, RAG, graph RAG, context engineering, model evaluation, fine-tuning, AI personalization, recommendation systems, distributed systems, large-scale systems (8B+ QPS), platform & developer tooling, privacy-aware data engineering, machine learning, technical leadership.


## Tech Stack

**Languages:** Python, Rust, C++, TypeScript / JavaScript (Node.js), PHP / Hack, SQL

**AI / ML:** PyTorch, LLMs, RAG & graph RAG, AWS Bedrock, LangChain, vLLM, vector databases (pgvector, Pinecone), embeddings, model evaluation & fine-tuning, Model Context Protocol (MCP)

**Infrastructure & Cloud:** AWS, Kubernetes, Docker, Terraform, gRPC, distributed systems

**Data:** PostgreSQL, Kafka, Redis, large-scale data pipelines


## Projects

### Property questions and answers
[booking.com/hotel/nl/luxury-suites-amsterdam](https://booking.com/hotel/nl/luxury-suites-amsterdam)

Introduced a major new feature, a questions-and-answers product, into Booking.com's core funnel, helping users make better-informed accommodation choices for their trips while also benefiting Booking.com.

### Enduro.js
[github.com/Gottwik/Enduro](https://github.com/Gottwik/Enduro) · ★ 681

Designed and open-sourced a lightweight, code-only content management system, built for developer speed and simplicity over heavyweight, database-driven alternatives.


## Education

### Master's Degree in Aerospace Engineering
University of Zilina
*2010 - 2012*

### Bachelor's Degree in Aerospace Engineering
University of Zilina
*2007 - 2010*