Dev Patel

Welcome

Dev Patel

I'm an engineer working on ML systems, post-training infrastructure, health-tech, and venture-backed startups.

LinkedInXEmailGithub

Now

I’m interested in building systems where technology, infrastructure, and markets reinforce each other. I care deeply about building technically ambitious products and organizations with long-term leverage, strong execution, and durable advantages.

Lately, I’ve been interested in how post-training systems, capex for infrastructure, and heterogeneous compute will reshape company building and the world. I’m also fascinated by the intersection of AI and life sciences, particularly by how new manufacturing paradigms and foundation models can change the way biology is understood, developed, and commercialized.

Roles

Summer 2026

Thrive Capital

Fellow

Selected as one of three fellows to work across portfolio impact, research, and incubations.

2025 - present

Sky Labs - NovaSky Group

Research Engineer

Contributing to SkyRL infrastructure for large-scale post-training, including training-inference mismatch, distributed training, and RL optimizations.

Summer 2025

AMD

AI Software Developer Intern

Worked on AI compiler and heterogeneous compute for Ryzen AI, including LLVM optimizations, NPU profiling, and PyTorch/ONNX workflows.

2024 - 2025

BAIR

Applied ML Researcher

Building language-model-based AFib risk prediction systems over large-scale EHR data and patient trajectory representations.

Summer 2024

Gladstone Institutes, Ye Lab

Deep Learning Intern

Built genomic foundation models and adapted protein models for variant functionalization and multiomics representation learning.

Writing

Venture EDU

A lecture-note version of a venture diligence deck: what to look for in healthtech startups, why business models matter, and how venture itself is changing.

Projects

2026

SkyRL

Contributing to SkyRL, an open-source modular reinforcement learning framework for LLM post-training, with work across rollout generation, Megatron parallelism, vLLM/SGLang integration, GPU CI, and distributed RL training infrastructure.

RLLLM post-trainingMegatronvLLMdistributed systems
2025

AFib Risk Prediction

Large-scale AFib risk prediction and patient trajectory modeling using longitudinal EHR data, deep learning, and medical foundation models at BAIR and Computational Precision Health.

EHRclinical MLPySparklanguage models
2025

MicroML

Built a lightweight deep learning framework in modern C++ with automatic differentiation, SIMD-optimized tensor kernels, and end-to-end neural network training infrastructure.

C++20autogradSIMDML systems
2025

Introspect AI

Built a 3D semantic visualization system for exploring AI conversation histories using embeddings, UMAP, clustering, and interactive WebGL interfaces.

AI toolingembeddingsThree.jssemantic search