GoodInternship

PyTorch in internship postings

From the 11,714 open internship postings we track that have a description · as of Nov 11, 2026

Postings that ask for it2.4%277 postings (283 requisitions) of 11,714 postings with a description
Remote7.6%21 of the 277 PyTorch postings
Median pay with it+103.7%$55.00/h (N=102) vs $27.00/h for all postings with pay (N=4377)

How to read this: this compares pay in postings that ask for the skill with pay in all postings. It mostly shows which kinds of jobs ask for it — not what the skill alone does to your pay.

Fields that ask for it
Share of the 277 PyTorch postings in each field. A posting can be in several fields, or none.
  1. Data Science68.2% (189 of 277)
  2. Computer Science18.1% (50 of 277)
  3. Engineering13.4% (37 of 277)
  4. Business6.9% (19 of 277)
  5. Product & Design2.5% (7 of 277)
  6. Finance1.1% (3 of 277)
Employers asking for it most
Postings per employer, out of the 277 PyTorch postings.
  1. NVIDIA8.7% (24 of 277)
  2. General Motors6.9% (19 of 277)
  3. Toyota Research Institute4.7% (13 of 277)
  4. Microsoft2.9% (8 of 277)
  5. Qualcomm2.9% (8 of 277)
  6. JPMorgan Chase1.8% (5 of 277)
  7. Bedrock Robotics1.8% (5 of 277)
  8. Merck1.4% (4 of 277)
Learned PyTorch? Add one of these next
The skills PyTorch postings most often ask for alongside it: the share of PyTorch postings that also name each one.
  1. Python89.2% (247 of 277)
  2. TensorFlow54.9% (152 of 277)
  3. C++34.7% (96 of 277)
  4. Java14.1% (39 of 277)
  5. Git13% (36 of 277)
Open postings that ask for PyTorch
The newest ones, still open when these numbers were computed.

How these numbers are made

Counting unit: a posting is one role. Repeat requisitions from the same employer with the same title and the same location (city, state or province, country, and remote or not) count once, using the newest one. Counted over open, de-duplicated postings with a description (11,714), once per posting; the 277 PyTorch postings stand for 283 requisitions. Every share, median and pairing below uses postings, not requisitions. A posting counts as asking for PyTorch when its description matches: word, any case. A skill is never counted for its own employer's postings. Shares show their N and denominator: 2.4% (277 of 11,714).

Pay compares hourly US-dollar pay (weekly, monthly and yearly pay converted, ranges at their midpoint), only for postings that publish it: 102 PyTorch postings against 4377 postings overall. Medians, not averages.