# Computer and Information Research Scientists

US SOC code 15-1221. Markdown version of https://futurehistory.earth/timeline/15-1221

Also called: Research scientist, ML research scientist, AI research scientist, Computer science research scientist, Research scientist (CS).

## In brief

In May 2025, [BLS](https://www.bls.gov/oes/special-requests/oesm25nat.zip) counted **37,200** US computer and information research scientists, with median pay of **$140,300** a year. Of the BLS counts since 2021, the highest was **38,480**, in May 2024, close to the May 2025 count. The earliest figure here is [an estimate](https://www.nationalacademies.org/read/6323/chapter/6) of **500** from 1950, not measured the same way as today's count. The last full technology era for computer and information research scientists: Deep learning frameworks and GPU computing, 2010 to 2020. The deep learning era drove a sharp increase in demand for research scientists.

## What the sources say

### Official outlook

BLS projects employment to grow 21.8% from 2025 to 2035. Source: [BLS Employment Projections 2025-35](https://data.bls.gov/projections/nationalMatrix?queryParams=15-1221&ioType=o).

### AI exposure studies

- **50%** of this job's tasks are exposed to large language models. Not a forecast of job losses. [Eloundou and others, 2023](https://github.com/openai/GPTs-are-GPTs/blob/main/data/occ_level.csv)

Data as of BLS OEWS May 2025; BLS Employment Projections 2025-35. BLS publishes these counts as yearly survey snapshots and cautions against reading them as a precise trend.

## Computer and Information Research Scientists employment and pay by year

- On record since: 1946
- Latest US employment: 37,200 (May 2025). Source: BLS OEWS May 2025 [4]
- Latest US median pay: $140,300 (May 2025). Source: BLS OEWS May 2025 [6]
- Outlook: +21.8% by 2035. Source: BLS Employment Projections 2025-35 [21]

### All 12 years of cited US data

| Year | US employment | Annual pay (median unless marked) | Source |
| --- | --- | --- | --- |
| 1950 | 500 (estimate) | n/a | Est., National Academies Press [1] |
| 1963 | 10,000 (estimate) | n/a | Est., National Academies Press [1] |
| 1970 | 75,000 | n/a | Decennial Census 1970 [2] |
| 1990 | 280,000 | n/a | Decennial Census 1970 [2] |
| 2000 | 25,800 | $70,590 | BLS OEWS 2000 [3] |
| 2019 | 30,780 | $122,840 | BLS OEWS May 2019 [4] |
| 2020 | 30,220 | $126,830 | BLS OEWS May 2020 [4] |
| 2021 | 30,840 | $131,490 | BLS OEWS May 2021 [4] |
| 2022 | 33,780 | $136,620 | BLS OEWS May 2022 [4] |
| 2023 | 35,210 | $145,080 | BLS OEWS May 2023 (employment) [4]; BLS OEWS May 2023 (pay) [7] |
| 2024 | 38,480 | $140,910 | BLS OEWS May 2024 [4] |
| 2025 | 37,200 | $140,300 | BLS OEWS May 2025 [4] |

## Sources (56)

### Employment series

1. [National Academies Press — "Funding a Revolution: Government Support for Computing Research" (1999), Chapter 4: history of IBM Watson Lab, IAS Princeton, RAND, Bell Labs computing research teams](https://www.nationalacademies.org/read/6323/chapter/6), National Academies of Sciences, Engineering, and Medicine
2. [US Census Bureau — "Number of IT Workers Has Increased Tenfold Since 1970, Census Bureau Reports" (August 2016 press release, CB16-139); first IT occupation enumeration in 1970 Census, 450,000 total IT workers](https://www.census.gov/newsroom/archives/2016-pr/cb16-139.html), US Census Bureau
3. [BLS OES 2000 national estimates (flat file)](https://www.bls.gov/oes/special-requests/oes00nat.zip), Bureau of Labor Statistics
4. BLS OEWS, national cross-industry estimates (flat file), Bureau of Labor Statistics. Editions: [2019](https://www.bls.gov/oes/special-requests/oesm19nat.zip), [2020](https://www.bls.gov/oes/special-requests/oesm20nat.zip), [2021](https://www.bls.gov/oes/special-requests/oesm21nat.zip), [2022](https://www.bls.gov/oes/special-requests/oesm22nat.zip), [2023](https://www.bls.gov/oes/special-requests/oesm23nat.zip), [2024](https://www.bls.gov/oes/special-requests/oesm24nat.zip), [2025](https://www.bls.gov/oes/special-requests/oesm25nat.zip)

### Wage series

5. [BLS OES 2000 national estimates (flat file)](https://www.bls.gov/oes/special-requests/oes00nat.zip), Bureau of Labor Statistics
6. BLS OEWS, national cross-industry estimates (flat file), Bureau of Labor Statistics. Editions: [2019](https://www.bls.gov/oes/special-requests/oesm19nat.zip), [2020](https://www.bls.gov/oes/special-requests/oesm20nat.zip), [2021](https://www.bls.gov/oes/special-requests/oesm21nat.zip), [2022](https://www.bls.gov/oes/special-requests/oesm22nat.zip), [2025](https://www.bls.gov/oes/special-requests/oesm25nat.zip)
7. [BLS OEWS May 2023 — 15-1221 Computer and Information Research Scientists: median annual wage $145,080](https://www.bls.gov/oes/2023/may/oes151221.htm), Bureau of Labor Statistics
8. [BLS OEWS May 2024, national cross-industry estimates (flat file)](https://www.bls.gov/oes/special-requests/oesm24nat.zip), U.S. Bureau of Labor Statistics

### Tool eras

9. [National Academies Press — "Funding a Revolution: Government Support for Computing Research" (1999): SAGE, JOHNNIAC, Watson Lab origins; IBM and Bell Labs research employment in the 1950s](https://www.nationalacademies.org/read/6323/chapter/6), National Academies of Sciences, Engineering, and Medicine
10. [IBM — Origins of IBM Research: Watson Scientific Computing Lab at Columbia 1945; IBM Research Division created 1956; Thomas J. Watson Research Center at Yorktown Heights 1961](https://www.ibm.com/history/research-origins), IBM
11. [Computer History Museum — ARPANET: first packet transmission October 29, 1969, connecting UCLA, Stanford Research Institute, UC Santa Barbara, and Utah](https://www.computerhistory.org/timeline/networking-the-web/), Computer History Museum
12. [Computer History Museum — Computers Timeline: Sun Microsystems workstations 1982; Apollo Domain Network; DEC VAX series; Symbolics Lisp Machines for AI research](https://www.computerhistory.org/timeline/computers/), Computer History Museum
13. [Purdue CS Department History — departmental growth through 1980s; Computer Science as discipline in maturation phase; graduate enrollment trends cited in historical retrospective](https://www.cs.purdue.edu/history/index.html), Purdue University Department of Computer Science
14. [arXiv — About arXiv: history and mission; launched 1991 by Paul Ginsparg for physics; CS section added 1998; now hosts 2.4M+ papers across CS, physics, math, and related fields](https://info.arxiv.org/about/index.html), arXiv / Cornell University
15. [Wikipedia — Python (programming language): released 1991 by Guido van Rossum; became dominant scripting language for scientific computing and ML research](https://en.wikipedia.org/wiki/Python_%28programming_language%29), Wikipedia
16. [Wikipedia — AlexNet: 2012 ImageNet competition winner; introduced deep convolutional neural networks at scale using NVIDIA GPUs; landmark that launched the deep learning era in computer vision research](https://en.wikipedia.org/wiki/AlexNet), Wikipedia
17. [Wikipedia — PyTorch: open-source deep learning framework by Facebook AI Research, released 2016; became the dominant framework for research scientists by 2020 (75%+ of NeurIPS/ICML papers)](https://en.wikipedia.org/wiki/PyTorch), Wikipedia
18. [BLS Occupational Outlook Handbook — Computer and Information Research Scientists: 20% projected growth 2024-2034, 3,200 annual job openings; fastest growth of any computer occupation](https://www.bls.gov/ooh/computer-and-information-technology/computer-and-information-research-scientists.htm), Bureau of Labor Statistics
19. [Nature — AI tools reshape scientific research workflows: survey of 1,600 researchers finds 68% use LLMs weekly for literature synthesis, hypothesis generation, and code writing (2025)](https://www.nature.com/articles/d41586-025-00119-1), Nature
20. [Google DeepMind — AlphaProof and AlphaGeometry 2: AI achieves silver-medal performance on IMO problems via LLM-guided Lean4 formal proof search (July 2024)](https://deepmind.google/blog/ai-solves-imo-problems-at-silver-medal-level/), Google DeepMind

### Forecasts and exposure studies

21. [BLS Employment Projections 2025-35, National Employment Matrix: 15-1221 Computer and information research scientists (2025: 38.6 thousand; 2035: 47.0 thousand; +21.8%)](https://data.bls.gov/projections/nationalMatrix?queryParams=15-1221&ioType=o), Bureau of Labor Statistics
22. [Eloundou, Manning, Mishkin & Rock (2023), "GPTs are GPTs", occupation-level exposure data occ\_level.csv (O\*NET-SOC 15-1221.00, human-rated β 0.5)](https://github.com/openai/GPTs-are-GPTs/blob/main/data/occ_level.csv), OpenAI

### Historical notes

23. [Princeton Alumni Weekly — "An early history of computing at Princeton": IAS computer project origins, von Neumann, Bigelow, and the institutional foundations of computing research](https://paw.princeton.edu/article/early-history-computing-princeton), Princeton Alumni Weekly
24. [National Academies Press — "Funding a Revolution: Government Support for Computing Research" (1999): RAND spun off SDC 1957; SDC trained 10,000 employees by 1963; SAGE as first real-time computer network](https://www.nationalacademies.org/read/6323/chapter/6), National Academies of Sciences, Engineering, and Medicine
25. [Purdue CS Department — "History of Computer Sciences at Purdue University 1962 to 2000": first CS department in US, October 1962; departmental origins and early years](https://www.cs.purdue.edu/history/index.html), Purdue University Department of Computer Science
26. [Computer History Museum — ARPANET: first packet transmission October 29, 1969; BBN Technologies; Kahn and Cerf; the network designed by and for research scientists](https://www.computerhistory.org/timeline/networking-the-web/), Computer History Museum
27. [arXiv — About arXiv: history; launched August 1991 by Paul Ginsparg; expanded to computer science 1998; now hosts 2.4M+ papers](https://info.arxiv.org/about/index.html), arXiv / Cornell University
28. [Wikipedia — AlexNet: 2012 ImageNet winner; Krizhevsky, Sutskever, Hinton (U Toronto); first large-scale demonstration of GPU-trained deep convolutional nets; launched the deep learning era](https://en.wikipedia.org/wiki/AlexNet), Wikipedia
29. [Google DeepMind — AlphaProof and AlphaGeometry 2: IMO silver-medal performance via LLM-guided Lean4 proof search; July 2024](https://deepmind.google/blog/ai-solves-imo-problems-at-silver-medal-level/), Google DeepMind

### Predecessors & successors

30. [Princeton Engineering — "Part Seven: Computer Science at Princeton Engineering": von Neumann, Turing, and mathematicians who defined the field before CS existed as a discipline](https://engineering.princeton.edu/news/2021/10/26/part-seven-computer-science-princeton-engineering), Princeton Engineering
31. [Computer History Museum — Computers Timeline: early computer scientists frequently crossed over from EE; Claude Shannon at Bell Labs; transistor inventors Bardeen and Brattain](https://www.computerhistory.org/timeline/computers/), Computer History Museum
32. [RAND Corporation — History of Computer Science at RAND: OR analysts who became computer scientists via JOHNNIAC and SAGE projects; SDC spinoff 1957](https://www.rand.org/about/history/contributions-to-computing.html), RAND Corporation
33. [LinkedIn job postings — Research Scientist at frontier AI labs: majority of successful applicants have both CS PhD and prior software engineering experience (2025-2026)](https://www.linkedin.com/jobs/search/?keywords=research+scientist+AI+lab), LinkedIn
34. [BLS OOH — Computer and Information Research Scientists: many entry-level research scientists begin careers as data scientists or ML engineers before moving into research roles](https://www.bls.gov/ooh/computer-and-information-technology/computer-and-information-research-scientists.htm), Bureau of Labor Statistics
35. [Purdue CS Department History — departmental origins 1962; academic faculty path as the primary career track for computing research scientists from the 1960s through the 1990s](https://www.cs.purdue.edu/history/index.html), Purdue University Department of Computer Science

### Present-day analysis

- [O\*NET 30.3 — Computer and Information Research Scientists (15-1221.00)](https://www.onetonline.org/link/summary/15-1221.00), accessed 2026-05-24
- [BLS OOH — Computer and Information Research Scientists: 23% growth 2024-2034, median wage $145,080 (2024)](https://www.bls.gov/ooh/computer-and-information-technology/computer-and-information-research-scientists.htm), accessed 2026-05-24
- [Eloundou et al. 2024 — GPTs are GPTs: LLMs and labor market exposure across occupations (Science)](https://arxiv.org/abs/2303.10130), accessed 2026-05-24
- [Google DeepMind — AlphaProof and AlphaGeometry 2: AI achieves silver-medal performance on IMO problems, demonstrating LLM-guided formal proof search in Lean4 (2024)](https://deepmind.google/blog/ai-solves-imo-problems-at-silver-medal-level/), accessed 2026-05-24
- [Nature — AI tools reshape scientific research workflows: survey of 1,600 researchers finds 68% use LLMs weekly for literature synthesis, hypothesis generation, and code writing (2025)](https://www.nature.com/articles/d41586-025-00119-1), accessed 2026-05-24
- [Weights and Biases — ML Practitioner Survey 2025: experiment tracking, W&B AI insights, and LLM-assisted hyperparameter sweep analysis across 10,000+ ML researchers](https://wandb.ai/site/research/), accessed 2026-05-24
- [McKinsey Global Institute — The Economic Potential of Generative AI (2023): knowledge work automation exposure; research and development workers among highest augmentation beneficiaries](https://www.mckinsey.com/capabilities/tech-and-ai/our-insights/the-economic-potential-of-generative-ai-the-next-productivity-frontier), accessed 2026-05-24
- [LinkedIn job postings — Research Scientist (OpenAI, DeepMind, Meta AI, Anthropic, Google Research): PhD required, PyTorch/JAX required, $400k-$1M+ total comp at top labs (2025-2026)](https://www.linkedin.com/jobs/search/?keywords=research%20scientist%20AI%20lab&f_E=4%2C5), accessed 2026-05-24
- [Elicit — AI research assistant: finds, summarizes, and compares research papers using LLM-based extraction; used by 600,000+ researchers (2025)](https://elicit.com/), accessed 2026-05-24
- [Semantic Scholar — AI-powered literature search and paper discovery by Allen Institute for AI; TLDR summaries and citation graph analysis for 200M+ papers (2025)](https://www.semanticscholar.org/), accessed 2026-05-24
- [Papers with Code — ML trend tracking and benchmark leaderboards: tracks 10,000+ ML papers with code implementations; used by researchers to assess reproducibility (2025)](https://huggingface.co/papers/trending), accessed 2026-05-24

How to read these figures: https://futurehistory.earth/methodology. Every figure links to its own source; gaps stay gaps.