Overview
Software Engineering Intern @Options Technology
Engineering Manager & ML Engineer @Generative AI at Cornell
GitHub Contributions
Hello
- I’m Mateo — an Electrical & Computer Engineering student at Cornell, working on the infrastructure that serves large models and the research that stress-tests them.
- At Options Technology I build PrivateMind, private AI infrastructure serving open-weight models on B300/B200 GPUs across colocated data centers in New York and London.
- At the Cornell NLP Group I work on LLM safety alignment, evaluating frontier models under adversarial prompting.
- Author of Open Weight, Open Risk, a training-free jailbreak that steers a target model’s chain-of-thought, and Dispatch, an agentic pentesting platform that won Best Developer Tool at the 2026 Cornell AI Hackathon.
Blog(1)
Stack
Infrastructure
Metrics
Experience
Options Technology
- Location
- New York, New York
- Location type
- (On-site)
- Employment status
- Current
- Developing PrivateMind, Options IT's secure and private AI infrastructure for enterprise environments.
- Deploying distributed systems to serve open-weight models across NVIDIA B300/B200 GPU infrastructure.
- Leveraging data center colocation across major financial hubs New York and London to deliver low-latency compute while enforcing hardware-isolated, zero-trust boundaries for T1 financial institutions.
- Kubernetes
- OpenShift
- vLLM
- SGLang
- CUDA C++
- Distributed Systems
Cornell University
- Location
- Ithaca, New York
- Location type
- (On-site)
- Led 8-person engineering team at Cornell in building an ESG risk monitoring platform for Investcorp's investment portfolio business, driving engineering efforts while coordinating with stakeholders worldwide (Feb. 2026 – May 2026).
- Implemented vendor due diligence agentic pipeline with fintech company QuickFi (Sept. 2025 – Dec. 2025).
- TypeScript
- Python
- Agentic Pipelines
- Engineering Management
- Collaborated with a PhD researcher in Cornell NLP Group on LLM safety alignment.
- Engineered pipeline to evaluate safety-alignment across frontier models in adversarial prompt settings.
- Worked with B200-class GPUs to run LLMs locally, using tools like vLLM and SGLang.
- Python
- vLLM
- SGLang
- LLM Safety
- Evaluation
Hospital for Special Surgery
- Location
- New York, New York
- Location type
- (On-site)
- Python
- ESM
- Protein Embeddings
- R
Education
- Bachelor of Science in Electrical & Computer Engineering, GPA 3.75 / 4.00.
- Expected graduation: December 2028.
- C
- C++
- CUDA C++
- Python
- Rust
- Computer Architecture
- Machine Learning
Projects(4)
Agentic penetration testing platform that turns vulnerability findings into ready-to-merge GitHub PRs. / Best Developer Tool — Cornell AI Hackathon 2026
- Orchestrates security agents using Mastra and OpenRouter, with isolated execution via Blaxel Sandboxes
- Slackbot and Datadog middleware for triggering scans and ingesting logs to surface findings
- Converts raw scan output into reviewable, mergeable remediation PRs
- TypeScript
- Mastra
- OpenRouter
- Blaxel Sandboxes
- Datadog
- Slack API
- Agentic Systems
- Security
Training-free jailbreak in which an uncensored model steers the target's chain-of-thought. / Publication · Source
- Achieves near-universal compliance across open-weight models (DeepSeek, Kimi, GLM, Qwen)
- Iterative prompt injection and LLM-as-judge pipelines to streamline evaluation
- Evaluated 15 frontier models across 800 WMDP-derived biosecurity and chemical security requests
- Python
- vLLM
- SGLang
- LLM Safety
- Red Teaming
- Evaluation
- Research
Awards(3)
Bookmarks(2)
The Techno-Optimist Manifesto
- Author
- Marc Andreessen
- Category
- Article
- Bookmarked on
The Bitter Lesson
- Author
- Rich Sutton
- Category
- Article
- Bookmarked on