Vai al contenuto
Andrej on GitHub

Andrej

@karpathy

I like to train Deep Neural Nets on large datasets.

  • Stanford
  • 491K
    Total stars
  • 223K
    Followers
  • 54
    Public repos
  • 878
    Contributions

8 following

Esporta e genera

Template resume README del profilo
Anteprima rapida
Condividi profilo

Profile tree

  • Contributions: 878. Past 12 months.
  • Followers: 222,902. People who follow you.
  • Total stars: 490,718. Across all repos.
  • Following: 8. People you follow.
  • Public repos: 54. Owned, non-fork.

A living snapshot

Every leaf is a headline number from Andrej’s public GitHub. The canopy fills out as the work grows - sway and all.

  • 878 Contributions Past 12 months
  • 223K Followers People who follow you
  • 491K Total stars Across all repos
  • 8 Following People you follow
  • 54 Public repos Owned, non-fork

Data from GitHub API · refreshed in real time

What your GitHub says

Based on public GitHub activity · how this is calculated

  • HTML is the largest language share across analyzed repositories, at 84%.

    High confidence

  • Public GitHub activity is implementation-heavy: commits account for 96% of tracked contribution activity.

    Medium confidence

  • Most analyzed repositories (29 of 40) show no public activity in over a year.

    Medium confidence

  • This analysis covers 40 of 54 public repositories (74%); broader totals may differ.

    High confidence

Profile health

Signals from analyzed public data

Activity 70

Developing

  • 878 contributions in the trailing year
Portfolio 35

Needs work

  • 3 of 40 analyzed repositories are actively maintained
  • 490,718 stars across analyzed repositories
Discoverability 39

Needs work

  • 1 of 40 analyzed repositories have no description
  • 38 of 40 have no topic tags
Maintenance 8

Needs work

  • 3 of 40 analyzed repositories were pushed to in the last 6 months
  • 29 show no public activity in over a year
Collaboration 6

Needs work

  • 7 pull requests and 6 code reviews in the trailing year
  • 165 repositories starred

Recruiter snapshot

A skim of public evidence, not a hiring judgement

Technical focus
HTML / Jupyter Notebook / Python / Cuda

Largest language shares across analyzed repositories

Project depth
3 of 40

Analyzed repositories rated established or maintained

Collaboration
7 PRs / 6 reviews

Public pull requests and code reviews in the trailing year

Community signals
490,718 stars / 72,108 forks

Summed across analyzed repositories

Recent activity
10 of last 12 months

Calendar months with at least one tracked contribution

Strongest signals

  • 878 contributions in the trailing year

878
Contributions
Past 12 months
0
Current streak
No active streak
15
Longest streak
Oct 3, 2025 – Oct 17, 2025
Sep Nov Dec Jan Feb Mar Apr May Jun Jul Aug Sep Mon Wed Fri
Less More

Language distribution for karpathy HTML 84.5%, Jupyter Notebook 8.0%, Python 3.6%, Cuda 1.5%, JavaScript 0.9%, C 0.6%, Other 0.9%
  • HTML 84.5%
  • Jupyter Notebook 8.0%
  • Python 3.6%
  • Cuda 1.5%
  • JavaScript 0.9%
  • C 0.6%
  • Other 0.9%

Frontend Backend Mobile DevOps Data AI Security Embedded Game Dev
  • Backend 100
  • Data 97
  • Frontend 50
  • DevOps 17
  • Embedded 10
  • AI 3
  • Mobile 0
  • Security 0
  • Game Dev 0

Longest streak
15d
Oct 2025
Favourite coding day
Saturday
113 contributions
Most active year
2024
1,254 contributions
Signature language
HTML
84% of tracked code
Flagship repository
autoresearch
96,657 stars
On GitHub since
2010
16 years and counting

Andrej is a developer on GitHub whose public work is written mostly in HTML, which accounts for roughly 84% of their committed code, alongside Jupyter Notebook and Python. Their account holds 54 public repositories, collecting 490,718 stars in total, led by autoresearch at 96,657. In the trailing year the account logged 878 contributions, including a longest run of 15 consecutive active days. Andrej has been on GitHub since 2010 and is followed by 222,902 developers.

This overview is generated from Andrej’s public GitHub activity at the time of the last refresh. Figures cover public repositories and the trailing twelve months of contributions.

  1. 01

    Add topics to nanoGPT

    High impact

    nanoGPT is pinned but has no topic tags, making its stack harder to identify.

    nanoGPT →
  2. 02

    Add topics to nanochat

    High impact

    nanochat is pinned but has no topic tags, making its stack harder to identify.

    nanochat →
  3. 03

    Review llm.c

    High impact

    llm.c is pinned to your profile but hasn't been pushed to in over a year.

    llm.c →
  4. 04

    Add topics to llm.c

    High impact

    llm.c is pinned but has no topic tags, making its stack harder to identify.

    llm.c →
  5. 05

    Review llama2.c

    High impact

    llama2.c is pinned to your profile but hasn't been pushed to in over a year.

    llama2.c →
  6. 06

    Add topics to llama2.c

    High impact

    llama2.c is pinned but has no topic tags, making its stack harder to identify.

    llama2.c →
  7. 07

    Add topics to micrograd

    High impact

    micrograd is pinned but has no topic tags, making its stack harder to identify.

    micrograd →

How these are calculated: see the methodology page .

AI agents running research on single-GPU nanochat training automatically

Python 96,657 ★ 13,521 forks 52 issues today

The simplest, fastest repository for training/finetuning medium-sized GPTs.

Python 63,341 ★ 10,889 forks 248 issues yesterday

The best ChatGPT that $100 can buy.

Python 58,238 ★ 8,158 forks 27 issues yesterday

LLM training in simple, raw C/CUDA

Cuda 31,047 ★ 3,766 forks 92 issues yesterday

LLM Council works together to answer your hardest questions

Python 24,966 ★ 4,366 forks 70 issues yesterday

A minimal PyTorch re-implementation of the OpenAI GPT (Generative Pretrained Transformer) training

Python 24,912 ★ 3,331 forks 49 issues yesterday

Inference Llama 2 in one file of pure C

C 20,106 ★ 2,635 forks 127 issues yesterday

A tiny scalar-valued autograd engine and a neural net library on top of it with PyTorch-like API

Jupyter Notebook 17,637 ★ 2,808 forks 31 issues yesterday

Most popular
autoresearch

96,657 stars

Most recently active
nanochat

Last pushed 16d ago

Longest maintained
karpathy.github.io

Active for 12.2y

Most discoverable
arxiv-sanity-lite

100% of discoverability signals present

nanoGPT

The simplest, fastest repository for training/finetuning medium-sized GPTs.

Python 63,341 ★ 10,889 forks
nanochat

The best ChatGPT that $100 can buy.

Python 58,238 ★ 8,158 forks
llm.c

LLM training in simple, raw C/CUDA

Cuda 31,047 ★ 3,766 forks
llama2.c

Inference Llama 2 in one file of pure C

C 20,106 ★ 2,635 forks
micrograd

A tiny scalar-valued autograd engine and a neural net library on top of it with PyTorch-like API

Jupyter Notebook 17,637 ★ 2,808 forks

  • autoresearch 96,657 ★

    AI agents running research on single-GPU nanochat training automatically

  • nanoGPT 63,341 ★

    The simplest, fastest repository for training/finetuning medium-sized GPTs.

  • nanochat 58,238 ★

    The best ChatGPT that $100 can buy.

  • LLM101n 37,488 ★

    LLM101n: Let's build a Storyteller

  • llm.c 31,047 ★

    LLM training in simple, raw C/CUDA

  • llm-council 24,966 ★

    LLM Council works together to answer your hardest questions

All repositories →

Public visibility is moderately spread across several repositories rather than one or two.

Top 1 repository 20% of stars
Top 3 repositories 44% of stars
Top 5 repositories 58% of stars

Leading repository: autoresearch (96,657 stars, 490,718 total across analyzed repositories)

Technology breakdown

Derived from repository languages and topics - horizontal bars, never a pie.

Languages

  • HTML 84%
  • Jupyter Notebook 8%
  • Python 4%
  • Cuda 2%
  • JavaScript 1%
  • C 1%

Frameworks

  • Flask 100%

README, contributing-guide and GitHub Actions badges below are detected by exact filename on each repository’s default branch - they report presence, not quality.

  • nanochat Maintained
    Popularity 60
    Discoverability 50
    Maintenance 100
    README No contributing guide No CI workflow
  • autoresearch Developing
    Popularity 100
    Discoverability 25
    Maintenance 40
    README No contributing guide No CI workflow
  • nanoGPT Developing
    Popularity 67
    Discoverability 50
    Maintenance 40
    README No contributing guide No CI workflow
  • micrograd Established
    Popularity 18
    Discoverability 50
    Maintenance 85
    README No contributing guide No CI workflow
  • Popularity 2
    Discoverability 100
    Maintenance 0
    README No contributing guide No CI workflow
  • llm.c Dormant
    Popularity 32
    Discoverability 50
    Maintenance 15
    README No contributing guide CI: 3 workflows
  • Popularity 2
    Discoverability 25
    Maintenance 65
    No README No contributing guide No CI workflow
  • llm-council Developing
    Popularity 26
    Discoverability 25
    Maintenance 40
    README No contributing guide No CI workflow
  • rustbpe Developing
    Popularity 1
    Discoverability 50
    Maintenance 40
    README No contributing guide CI: 2 workflows
  • Popularity 6
    Discoverability 75
    Maintenance 0
    README No contributing guide No CI workflow
  • minGPT Dormant
    Popularity 26
    Discoverability 50
    Maintenance 0
    README No contributing guide No CI workflow
  • Popularity 26
    Discoverability 50
    Maintenance 0
    README No contributing guide No CI workflow

Scarica resume PDF

I tuoi progetti e le tue competenze sono già compilati da GitHub. Aggiungi ciò che GitHub non pubblica - dettagli di contatto, esperienza ed educazione - ed esporta un PDF di una pagina compatibile con gli ATS.

Tutto ciò che scrivi rimane in questo browser (localStorage). Non viene caricato nulla.

Aggiungi i tuoi dati
Contatto
Esperienza
Istruzione

54
Public repos
40 analyzed
491K
Total stars
Across all repos
878
Contributions
Past 12 months
7
Pull requests
Opened this year
13
Public gists
Snippets shared
16 y
Account age
Since 2010
80 %
Profile score
Completeness

Repository explorer

Name Language Stars Forks Size Updated

Coding statistics

Largest repository
27.2 MB
convnetjs
Smallest repository
10 KB
rendergit
Oldest repository
Dec 2, 2011
researchpooler
Newest repository
Mar 14, 2026
jobs
Most starred
96,657 ★
autoresearch
Most forked
13,521 forks
autoresearch
Most watched
733 watchers
autoresearch
Longest description
281 characters
researchpooler
Most open issues
248 open
nanoGPT

Contribution patterns

By year

  • 2026
    320
  • 2025
    743
  • 2024
    1,254
  • 2023
    561
  • 2022
    153
  • 2021
    179

By day of week

  • Sun
    82
  • Mon
    111
  • Tue
    111
  • Wed
    66
  • Thu
    65
  • Fri
    91
  • Sat
    113

Best coding day: Sat

By month (last 12)

  • Oct
    191
  • Nov
    38
  • Dec
    50
  • Jan
    98
  • Feb
    81
  • Mar
    82
  • Apr
    33
  • May
    1
  • Jun
    0
  • Jul
    24
  • Aug
    1
  • Sep
    0

GitHub journey

  1. April 2010

    Joined GitHub

    Created the @karpathy account.

  2. December 2011

    Created first repository

    Published researchpooler, the oldest repository still on the account.

  3. December 2011

    First starred repository

    researchpooler is the earliest-created repository to have earned stars.

  4. May 2015

    Reached 10 repositories

    char-rnn was the 10th public repository created.

  5. June 2024

    Most productive year: 2024

    1,254 contributions recorded in 2024, more than any other year on record.

  6. March 2026

    Latest repository

    jobs is the most recently created repository.

Developer evolution

Your GitHub career, year by year.

2011

Started with Python.

Python
Repos this year
1
Total repos
1
Stars, as of today
506
researchpooler
2012

Shifted from Python to JavaScript.

JavaScript
Repos this year
1
Total repos
2
Stars, as of today
1.2k
svmjs
2014

Went deeper into JavaScript.

CSS HTML JavaScript Python
Repos this year
6
Total repos
8
Stars, as of today
23k
convnetjs
2015

Shifted from JavaScript to Jupyter Notebook.

HTML Jupyter Notebook Lua Python
Repos this year
5
Total repos
13
Stars, as of today
49k
char-rnn
2016

Shipped 1 repository.

Repos this year
1
Total repos
14
Stars, as of today
50k
paper-notes
2018

Started with Python.

Python
Repos this year
1
Total repos
15
Stars, as of today
50k
pytorch-made
2019

Shifted from Python to Jupyter Notebook.

Jupyter Notebook
Repos this year
1
Total repos
16
Stars, as of today
51k
pytorch-normalizing-flows
2020

Shifted from Jupyter Notebook to Python.

Jupyter Notebook Python
Repos this year
3
Total repos
19
Stars, as of today
94k
minGPT
2021

Shifted from Python to Jupyter Notebook.

Jupyter Notebook Python
Repos this year
3
Total repos
22
Stars, as of today
99k
Contributions
179
cryptos
2022

Went deeper into Jupyter Notebook.

Jupyter Notebook Python
Repos this year
4
Total repos
26
Stars, as of today
191k
Contributions
153
nanoGPT
2023

Shifted from Jupyter Notebook to C.

C Python
Repos this year
2
Total repos
28
Stars, as of today
217k
Contributions
561
llama2.c
2024

Shifted from C to Python.

Cuda Python
Repos this year
4
Total repos
32
Stars, as of today
301k
Contributions
1,254
LLM101n
2025

Went deeper into Python.

Python
Repos this year
5
Total repos
37
Stars, as of today
392k
Contributions
743
nanochat
2026

Shifted from Python to HTML.

HTML Python Rust
Repos this year
3
Total repos
40
Stars, as of today
491k
Contributions
320
autoresearch

Built from karpathy’s public, non-fork repositories. Star totals reflect current counts, not historical snapshots.

Achievements

7 / 9 unlocked

Explorer
Created repositories in 8 or more languages
Builder
More than 50 public repositories
Open Source Enthusiast
100+ stars received across all repositories
Consistent
A 100-day-or-longer contribution streak
Documentation Hero
5 or more repositories with a detailed description
Popular
500 or more followers
Prolific Committer
1,000 or more commits in the past 12 months
Long Timer
5 or more years on GitHub
Star Collector
Starred 100 or more repositories

Open source activity

490,718
Stars received
165
Repositories starred
72,108
Forks received
3
Issues opened
Past 12 months
7
Pull requests opened
Past 12 months
6
Reviews given
Past 12 months
13
Public gists

Language evolution

The primary languages of repositories created each year.

2011
  • Python
2012
  • JavaScript
2014
  • CSS
  • HTML
  • JavaScript
  • Python
2015
  • HTML
  • Jupyter Notebook
  • Lua
  • Python
2018
  • Python
2019
  • Jupyter Notebook
2020
  • Jupyter Notebook
  • Python
2021
  • Jupyter Notebook
  • Python
2022
  • Jupyter Notebook
  • Python
2023
  • C
  • Python
2024
  • Cuda
  • Python
2025
  • Python
2026
  • HTML
  • Python
  • Rust

Repository growth

Repositories created per year, and the running total.

1 repositories created in 2011 1 repositories created in 2012 6 repositories created in 2014 5 repositories created in 2015 1 repositories created in 2016 1 repositories created in 2018 1 repositories created in 2019 3 repositories created in 2020 3 repositories created in 2021 4 repositories created in 2022 2 repositories created in 2023 4 repositories created in 2024 5 repositories created in 2025 3 repositories created in 2026 1 total repositories by end of 2011 2 total repositories by end of 2012 8 total repositories by end of 2014 13 total repositories by end of 2015 14 total repositories by end of 2016 15 total repositories by end of 2018 16 total repositories by end of 2019 19 total repositories by end of 2020 22 total repositories by end of 2021 26 total repositories by end of 2022 28 total repositories by end of 2023 32 total repositories by end of 2024 37 total repositories by end of 2025 40 total repositories by end of 2026 2011 2012 2014 2015 2016 2018 2019 2020 2021 2022 2023 2024 2025 2026 40 0
Created that year Cumulative total

Language tree

Built automatically from repository languages and topics.

Programming Languages
  • HTML
  • Jupyter Notebook
  • Python
  • Cuda
  • JavaScript
  • C
Frontend & Backend
  • Flask 1