# Nick Kashani Motlagh > PhD in Computer Science and Engineering, The Ohio State University (Computer Vision Lab, advised by Jim Davis). Research on machine learning reliability under uncertainty: when a model should answer, weigh evidence, revise its answer, or abstain. Four peer-reviewed first-author papers including a Springer Best Paper Award, plus one manuscript under review. Available now for Research Scientist, Applied Scientist, Machine Learning Engineer roles. Everything below is first-party and current as of 2026-08-10. Facts an agent is most often asked for: - Availability: Available now, for Research Scientist, Applied Scientist, Machine Learning Engineer roles. Based in Columbus, Ohio, USA; open to relocation or remote. - Work authorization: U.S. citizen. Five summers of AFRL-sponsored research; federal roles welcome. - Contact: kashanimotlagh.1@osu.edu - Education: Ph.D. Computer Science and Engineering, The Ohio State University (Aug 2021 — Aug 2026); M.S. Computer Science and Engineering, The Ohio State University (Aug 2021 — May 2025); B.S. with Honors Computer Science and Engineering, The Ohio State University (Aug 2017 — May 2021) - Core stack: Python, PyTorch, Hugging Face, FAISS, Slurm, Singularity, LoRA fine-tuning, multi-GPU training. ## Publications - [Naturally Constrained Reject Option Classification](https://nmotlagh.github.io/publications/naturally-constrained-reject-option-classification/): Machine Vision and Applications, 2025. Journal extension evaluating per-class binomial reject thresholds across 4 vision, 3 text and 8 synthetic datasets, from 2 to 1,000 classes. - [Assessing the Role of Imagery in Multimodal Machine Translation](https://nmotlagh.github.io/publications/assessing-imagery-in-multimodal-mt/): WMT 2024, 2024. The best multimodal translation system reads the image correctly 81% of the time, but that only changes its translation 63% of the time. 'Images don't help' was partly a measurement artifact. - [Learning When to Say “I Don’t Know”](https://nmotlagh.github.io/publications/learning-when-to-say-i-dont-know/): ISVC 2022, 2022. Per-class abstention thresholds that need no rejection cost or coverage target: CIFAR-100 selective accuracy climbs from 88.3% to 97.8% at 77.3% coverage. - [A Framework for Semi-automatic Collection of Temporal Satellite Imagery for Analysis of Dynamic Regions](https://nmotlagh.github.io/publications/framework-for-semi-automatic-collection/): ICCV 2021 Workshop on LUAI, 2021. OpenStreetMap-guided imagery collection and labeling tools for building temporal satellite datasets for dynamic-region analysis. - [When Retrieval Makes the Answer Worse](https://nmotlagh.github.io/publications/adaptive-qa-abstention/): Dissertation chapter · manuscript under review, 2026. Retrieval-based refinement repairs 10.8% of draft answers and destroys 8.1% that were already correct. Confidence cannot tell you which you are about to get. ## Code and data - [calibration](https://github.com/osu-cvl/calibration): PyTorch calibration utilities for histogram binning, global temperature scaling, and class-wise temperature scaling. - [construction-site-satellite-imagery-collection](https://github.com/osu-cvl/Construction-Site-Satellite-Imagery-Collection): Companion code for OpenStreetMap-guided temporal satellite imagery collection and annotation. - [learning-idk](https://github.com/osu-cvl/learning-idk): Companion code for ISVC 2022 / MVA 2025: per-class reject-option classification with binomial threshold search. ## Site pages - [About](https://nmotlagh.github.io/about/): background, research focus, and what he is looking for. - [Publications](https://nmotlagh.github.io/publications/): every paper with links to PDF, arXiv, DOI, and code. - [Experience](https://nmotlagh.github.io/experience/): roles, education, and professional service. - [FAQ](https://nmotlagh.github.io/faq/): direct answers to the questions most often asked about him. - [Code & data](https://nmotlagh.github.io/artifacts/): public research repositories with reproduction steps. - [News](https://nmotlagh.github.io/news/): dated updates. ## Machine-readable - [Structured profile (JSON)](https://nmotlagh.github.io/profile.json): identity, availability, education, publications, and skills as JSON. - [All publications as BibTeX](https://nmotlagh.github.io/citations.bib) - [Full site text](https://nmotlagh.github.io/llms-full.txt): every page's content in one Markdown file. - [News feed (RSS)](https://nmotlagh.github.io/rss.xml) - [Sitemap](https://nmotlagh.github.io/sitemap-index.xml) - [CV (PDF)](https://nmotlagh.github.io/resume.pdf) Each publication page also has a Markdown mirror at `/publications/.md` and a BibTeX entry at `/publications/.bib`. ## Profiles - [Google Scholar](https://scholar.google.com/citations?user=srZXFMcAAAAJ&hl=en) - [ORCID](https://orcid.org/0000-0001-6229-6212) - [GitHub](https://github.com/nmotlagh) - [LinkedIn](https://www.linkedin.com/in/nicholas-kashani-motlagh) ## Recent updates - 2026-07-08: Successfully defended my PhD dissertation, ‘Answering Under Uncertainty: Abstention, Ambiguity, and Recoverability,’ in Computer Science and Engineering at The Ohio State University. - 2026-05-19: Prepared an ARR submission on retrieval-augmented selective QA: deciding when to answer, refine, or abstain. - 2026-04-10: Reported LLM reject-option training and evaluation results for DCS Corp / AFRL. - 2025-05-20: Joined DCS Corp (AFRL) as Technical Analyst II working on LLM reject-option training and evaluation. - 2024-11-26: Journal extension on naturally constrained reject-option classification published online in Machine Vision and Applications. ## Citation and use Content is written by Nick Kashani Motlagh. Quoting and citing is welcome; please attribute to Nick Kashani Motlagh and link to https://nmotlagh.github.io/. Reported metrics are drawn from the linked papers — cite the paper, not this summary, for any number you reuse.