Jyoti Aneja

Career

Aneja trained in physics before moving to machine learning, earning master's degrees in physics and then a PhD at the University of Illinois Urbana-Champaign in 2021, advised by Alex Schwing and working closely with David Forsyth. Aneja joined Microsoft Research in November 2021, with research at the intersection of computer vision, deep learning and natural language processing.¹²

Phi models

Aneja's site lists core contributions to Microsoft's Phi-1, Phi-2, Phi-3 and Phi-4 small language models, and Aneja appears on the author list of the December 2024 Phi-4 technical report. Aneja co-led Phi-4-reasoning-vision-15B, a compact open-weight multimodal reasoning model; the March 2026 Microsoft Research post by Aneja and colleagues describes its vision encoder and image-processing choices, data curation, and the use of a mixture of reasoning and non-reasoning training data.¹⁴³

Work and contributions

• Phi-4-reasoning-vision-15B Open weights

  Co-led the model and is first-listed author of its technical report and of the Microsoft Research post describing how it was trained.¹³

• Phi-4 Open weights

  Core contributor to Phi-4 and listed among the authors of the Phi-4 technical report.¹⁴

Timeline

1. 2021Completed a PhD at the University of Illinois Urbana-Champaign advised by Alex Schwing.¹
2. Nov 2021Joined Microsoft Research.²
3. Dec 2024Co-authored the Phi-4 technical report.⁴
4. Mar 4, 2026Published the Microsoft Research post on training Phi-4-reasoning-vision-15B as first-listed author.³

Where they work

• Microsoft 75 people in the wiki

Related people

• Xiaodong HeAlso at Microsoft
• Karén SimonyanAlso at Microsoft
• Hoifung PoonAlso at Microsoft
• Bin XiaoAlso at Microsoft
• Jianfeng GaoAlso at Microsoft
• Hassan ShapourianMultimodal models · Computer vision
• Ali FarhadiComputer vision · Multimodal models
• Matt DeitkeMultimodal models · Computer vision
• Aravind SrinivasLanguage models · Computer vision
• Cordelia SchmidComputer vision · Multimodal models