Skip to content
View danielfleischer's full-sized avatar
35 followers · 3 following

Sponsoring

@tarsius
@minad
@gnachman
@xenodium

Achievements

Achievement: YOLOAchievement: Pair ExtraordinaireAchievement: Public SponsorAchievement: Pull Sharkx3Achievement: QuickdrawAchievement: Starstruck

Achievements

Achievement: YOLOAchievement: Pair ExtraordinaireAchievement: Public SponsorAchievement: Pull Sharkx3Achievement: QuickdrawAchievement: Starstruck

Block or report danielfleischer

Block user

Prevent this user from interacting with your repositories and sending you notifications. Learn more about blocking users.

You must be logged in to block users.

Content in all repositories owned by your account will be closed.
Maximum 250 characters. Please don’t include any personal information such as legal names or email addresses. This note will only be visible to you.
Report abuse

Contact GitHub support about this user’s behavior. Learn more about reporting abuse.

Report abuse
danielfleischer/README.md

Daniel Fleischer 👋

I build LLM systems: agents, reinforcement learning for reasoning, retrieval, and the evaluation harnesses that keep them honest.

Intel Labs, 2022–2026:

  • fastRAG — efficient retrieval-augmented generation; late-interaction retrieval (ColBERT + PLAID), RAG on CPU.
  • RAG-FiT — fine-tuning LLMs for RAG, from dataset creation to evaluation. Paper.
  • DeepMath — GRPO training for tool-using math agents that run Python inside their own reasoning. Blog.
  • Xe-Forge — LLM agents that optimize GPU kernels, verified on real hardware. Paper.

Personal projects:

Website · Publications · LinkedIn

Pinned Loading

  1. IntelLabs/fastRAG IntelLabs/fastRAG Public archive

    Efficient Retrieval Augmentation and Generation Framework

    Python 1.8k 166

  2. IntelLabs/Xe-Forge IntelLabs/Xe-Forge Public

    Multi-stage LLM agent pipeline for optimizing AI kernels on Intel GPU — from analysis to autotuning.

    Python 29 9

  3. IntelLabs/RAG-FiT IntelLabs/RAG-FiT Public archive

    Framework for enhancing LLMs for RAG tasks using fine-tuning.

    Python 765 61

  4. mu4easy mu4easy Public

    mu4e + mbsync configuration for multiple email accounts in Emacs.

    Emacs Lisp 77 6

  5. IntelLabs/DeepMath IntelLabs/DeepMath Public archive

    Simple framework for training and evaluating math reasoning agents using local models, GRPO and vLLM.

    Python 52 4

  6. otzar otzar Public

    Local hybrid search over your own files — dense vectors, BM25, and metadata filters in one query. The same binary is an MCP server, so agents can search your notes.

    Clojure 4