Companies using Keras

57,872 companies in our data use Keras, and 11,318 of them name it in job posts they wrote. Hiring for it has slowed, though 8,610 job posts still named it from October 2025 to August 2026. AI Engineer roles mention it most.

Data as of November 9, 2026 · Markdown version

Which companies name Keras in the most job posts?

Among the companies with the most job posts naming Keras are JPMorganChase (667), Deloitte (650) and IBM (490). Of the five companies in the table, Apple has the most employees who list it on their profiles: 719.
Company Posts People Recent job post
JPMorganChase Finance and Insurance · first seen Jul 2022 667 517 "Solid coding level in Python programming language, with experience in leveraging available libraries, like Tensorflow, Keras, Pytorch, Scikit-learn, or others, to dedicated projects."Data Scientist Lead- Python, ML & Gen AI, Nov 5, 2026
Deloitte Professional Services · first seen Jul 2022 650 711 "Must have strong hands-on experience with programming languages like Python, Cuda and SQL, and frameworks such as TensorFlow, PyTorch and Keras"Gen AI Architect-Lead AI and Data Solutions Engineer II, Oct 23, 2026
IBM Information · first seen Jul 2022 490 580 "…proficiency in Python and experience with AI frameworks such as TensorFlow, PyTorch, Keras, or Hugging Face, to develop and deploy AI models."data scientist-artificial intelligence, Nov 1, 2026
Booz Allen Hamilton Professional Services · first seen Aug 2022 389 100 "Experience with machine learning packages and building and testing neural network architectures using TensorFlow, Keras, or PyTorch"Real-Time Software Engineer, Nov 6, 2026
Apple Scientific and Technical Services · first seen Aug 2022 363 719 "Experience with Spark, TensorFlow, Keras, and PyTorch a plus"Machine Learning Engineer, Apple Store Online, May 31, 2026
Showing 1 to 5 of 11,318 companies with job posts Sign up to see all companies using Keras ›

What else do we know about Keras?

Is hiring for Keras growing?

A smaller share of job posts mentioned Keras from June to August 2026 than from March to May: 3.6 in every 10,000 job posts, down from 4.3. Its busiest month was November 2025, at 5.5.
Share of job posts naming Keras by month, October 2025 to August 2026, in job posts in every 10,000 Moves from 5.1 in October 2025 to 3.4 in August 2026, in job posts in every 10,000. 6.3 4.8 3.2 1.6 0 3.4 Oct '25 Nov Dec Jan '26 Feb Mar Apr May Jun Jul Aug
Show the data
MonthJob posts naming KerasAll job postsPer 10,000
Oct 20257261,413,6735.1
Nov 20251,0061,822,1195.5
Dec 20256471,236,1035.2
Jan 20269701,828,0025.3
Feb 20269621,856,6475.2
Mar 20268251,850,7114.5
Apr 20267801,728,8864.5
May 20267201,821,5094.0
Jun 20267311,909,4963.8
Jul 20266271,815,1803.5
Aug 20266161,835,6983.4

How many companies named Keras in a job post for the first time in recent months?

From June to August 2026, 178 companies posted their first job naming Keras; 215 did from March to May. That brings the companies naming it for the first time since October 2025 to 857, through August 2026.
March to May 2026 215
June to August 2026 178
October 2025 to August 2026 857

Companies whose first job post naming it appeared in that period. Recruiters, staffing firms and job boards are left out.

How do we know which companies use Keras?

Employees' profiles account for 80% of the companies here. 11,318 companies name Keras in job posts they wrote. 4,763 show up in both job posts and employees' profiles.
How we know Companies Share
Job posts name Keras
11,318 20%
Employees' profiles list Keras
46,554 80%

57,872 companies. Each is counted once, under the strongest evidence we have for it.

What size are the companies that use Keras?

Keras skews toward smaller firms. 36% of its users are companies with fewer than 50 employees, against 22% on the typical tech page. 3,562 of its companies have more than 5,000 employees.

Share of its companies Typical tech page

ShareTypical
Under 50 employees
36% 22%
50 to 999
45% 38%
1,000 to 5,000
13% 17%
Over 5,000
6.2% 23%

Where are Keras companies based?

Indonesia is 5.8 times as common among Keras companies as on the typical tech page, at 3.8%. The United States leads overall with 32%. Its companies are based in 171 countries and territories.
ShareTypical
United States
32% 43%
India
9.8% 6.9%
Germany
5.9% 4.4%
United Kingdom
5.8% 6.4%
France
4.6% 4.6%
Indonesia
3.8% 0.64%

Companies in 171 countries and territories. Shares are of companies with a known headquarters country.

Show the data
CountryCompaniesShareTypical tech page
United States17,32232%43%
India5,3379.8%6.9%
Germany3,2075.9%4.4%
United Kingdom3,1945.8%6.4%
France2,5254.6%4.6%
Indonesia2,0583.8%0.64%

Which teams work with Keras?

Data science teams are by far the most common home for Keras, at 358 companies. Next come AI teams (147 companies) and machine learning teams (107).
Team Companies
Data science
358
AI
147
Machine learning
107
R&D
101
Engineering
71
Data
67

What else do Keras companies use?

The tool most tied to Keras is NLTK. 9.7% of Keras companies use it, 15 times the usual rate for their size. spaCy is 14 times as common as usual, at 11%.

Share of the 8,636 Keras companies counted here whose job posts also name the tool Typical for companies the same size

Tool Also name it Typical Times as common
NLTK 842 companies
9.7% 0.65% 15x
spaCy 918 companies
11% 0.77% 14x
ONNX 528 companies
6.1% 0.51% 12x
TensorRT 513 companies
5.9% 0.52% 12x
Tool Also name it Typical Times as common
BERT 827 companies
9.6% 0.84% 11x
TensorFlow 8,180 companies
95% 9% 11x
Kubeflow 1,265 companies
15% 1.4% 10x
PyTorch 7,547 companies
87% 8.8% 10x

Which companies started naming Keras in job posts recently?

192 companies with 50 or more employees started naming Keras in job posts between April and August 2026. 23 of them have more than 5,000 employees. Another 5 were AI-native companies, which we list separately.
AI native any size · 5 new
Company Posts
Viven 1
Mid-market 1,000 to 5,000 employees · 53 new
Company Posts
Enterprise over 5,000 employees · 23 new
Company Posts
Swiggy 6

Companies that named Keras in a job post for the first time between April and August 2026, ranked by their job posts in those months. AI-native companies get a group of their own, whatever their size. Recruiters and job boards are left out.

Which roles mention Keras most?

AI Engineer and Data Scientist roles account for 51% of the job posts that name Keras. The top four roles cover 72% of these job posts.
Role Share Posts
AI Engineer
29% 2,471
Data Scientist
22% 1,902
Machine Learning
14% 1,218
Software Engineer
6.9% 597
Researcher
1.5% 128
MLOps Engineer
1% 89
Show the data
RoleShareOct 2025 to Aug 2026Mar to May 2026Jun to Aug 2026
AI Engineer29%2,471670553
Data Scientist22%1,902509440
Machine Learning14%1,218268247
Software Engineer6.9%597149120
Researcher1.5%1282824
MLOps Engineer1%892215

Which project themes come up in Keras job posts?

Teams most often pair Keras with Generative AI (4,229 job posts), then broad digital transformation, then Cloud Cost Optimization and Microservices Migration.
Project theme Job posts
Generative AI 4,229
Digital Transformation 2,585
Cloud Cost Optimization 288
Microservices Migration 193
About this data

Sumble built this page. Sumble gives sales, marketing and other GTM teams technographics and hiring signals for the companies they sell to: the technologies each company uses, which teams are hiring and for what, and who reports to whom. That helps a team know who to contact, why now, and what to say.

Sumble covers more than 3.6 million companies, 93 million job posts collected since 2022 and 194 million employee profiles, and tracks more than 37,000 technologies. It finds which technologies a company uses by reading its job posts and its employees' public profiles.

Sumble has a free plan, and the same data is available through its API and MCP server.

A company counts when its own job posts or its employees' public profiles name the technology. Hiring is that technology's share of all job posts collected each month. Quotes are copied word for word from the companies' job posts.

Every figure on this page comes from Sumble's own analysis, and you won't find it outside Sumble.