·
AI Infrastructure & Inference
Zach Nussbaum
Principal ML Engineer |Nomic AI
Co-creator of GPT4All and lead author of Nomic Embed, building fully open and reproducible embedding models and local LLM infrastructure.
Biography
Deep Research
Zach Nussbaum is a Principal Machine Learning Engineer at Nomic AI, where he leads the development of the Nomic Embed family of open-source text and vision embedding models. He is a co-creator of GPT4All, the pioneering open-source local LLM chatbot that became the third-fastest-growing GitHub repository of all time with over 77,000 stars and 250,000 monthly active users. Before Nomic, he worked at Deep Genomics on machine learning for drug discovery, contributing to BigRNA, a foundation model for tissue-specific RNA expression prediction. A former Division I baseball player at Davidson College, he also participated in open-community research with ML Collective and OpenBioML.
Local LLM InferenceOpen-Source Embedding ModelsReproducible ML TrainingMatryoshka Representation LearningMultimodal EmbeddingsMixture-of-Experts ArchitecturesDemocratizing AI AccessCode Retrieval & RankingComputational Biology & GenomicsWebGPU ML Acceleration
Timeline
12 Research12 total
2025
2024
2024-02Research
Appeared on MAD Podcast with Matt Turck: 'How Nomic AI Is Driving The Open Source Revolution'2024-02Research
Appeared on Weaviate Podcast: 'Matryoshka Embeddings' with Aditya Kusupati and Zain Hasan2024-06Research
Co-authored Nomic Embed Vision — multimodal embedding model sharing latent space with Nomic Embed Text2024-07Research
Published blog post on multimodal embedding observability2023
2021
2021-01Research
Presented 'A Tale of Two Long Tails' at ICML UDL WorkshopKey Contributions
GPT4All
Co-created the open-source local LLM chatbot that became the 3rd fastest-growing GitHub repo of all time (77k+ stars, 250k+ MAU). Demonstrated that capable assistant-style chatbots could run on consumer CPUs without API calls.
Nomic Embed
Lead author of the first fully reproducible, open-source, open-data text embedding model (v1) that outperformed OpenAI Ada-002. Evolved into v1.5 with Matryoshka dimensions, v2 with MoE architecture supporting ~100 languages, and a vision variant.
Nomic Embed Vision
Co-authored multimodal vision embedding model that shares the same latent space as Nomic Embed Text, enabling unified text-image retrieval with fully open weights and training code.
CoRNStack
High-quality contrastive dataset for code retrieval across multiple programming languages, accepted at ICLR 2025. Powers nomic-embed-code and CodeRankEmbed models.
Contrastors
Open-source PyTorch library for training contrastive models, used as the training framework behind the Nomic Embed model family.
BigRNA (Deep Genomics)
Contributed to training BigRNA, an RNA foundation model predicting tissue-specific RNA expression, splicing, microRNA sites, and RNA binding protein specificity from DNA sequence.
Notable Quotes
“
We describe the training of nomic-embed-text-v1, the first fully reproducible, open-source, open-weights, open-data, 8192 context length English text embedding model that outperforms both OpenAI Ada-002 and OpenAI text-embedding-3-small.
“
Large language models have recently achieved human-level performance on a range of professional and academic benchmarks. The accessibility of these models has lagged behind their performance.
“
In contrast with other open-source models, we release the full curated training data and code to allow for full replication of nomic-embed-text-v1. This sets a new standard for transparency in the embedding model space.
“
GPT4All aims to democratize access to LLMs. Quantized 4-bit versions of the model allow virtually anyone to run the model on CPU.
12 sources(click to expand)
Zach Nussbaum — personal websitezanussbaum — GitHub profileGPT4All: An Ecosystem of Open Source Compressed Language Models (EMNLP 2023)GPT4All Technical Report (Mar 2023)Nomic Embed: Training a Reproducible Long Context Text Embedder (arXiv)Nomic Embed Vision: Expanding the Latent Space (arXiv)Training Sparse Mixture Of Experts Text Embedding Models (arXiv)nomic-ai/gpt4all — GitHub repository (77k+ stars)GPT4All — Nomic product pageMatryoshka Embeddings — Weaviate Podcast (Feb 2024)How Nomic AI Is Driving The Open Source Revolution — MAD Podcast (Feb 2024)CoRNStack — ICLR 2025 dataset for code retrieval
Research generated March 19, 2026
AI Infrastructure & Inference/Zach Nussbaum
All Profiles