Technical Writer

Help Pangram explain fast-moving detection research with rigor and clarity.
About Pangram Labs

Pangram Labs is a seed-stage AI research lab building advanced detection products for text and images. We believe people and institutions need better ways to understand authenticity and authorship as generative AI becomes more capable and more widely used. Our team works on difficult detection problems and turns that research into products for customers across education, media, and enterprise settings.

As Pangram’s Technical Writer, you will own the written layer between our research, product, and the people who rely on them. You will turn experiments, model releases, studies, and product behavior into accurate materials for researchers, developers, institutional customers, and everyday users.

The ideal candidate should be familiar with machine learning and statistics concepts like model architecture, training data, benchmark design, uncertainty, and error rates. You should also possess excellent writing skills across a variety of domains, from dense technical documents to mathematical formulas to plain-language. You will be expected to shape documents for audiences with very different levels of technical depth, working directly with machine learning researchers, engineers, data researchers, and programmers, while serving as a consistent editorial owner across Pangram’s public and internal technical writing.

• Make Pangram’s methods, performance, and limitations understandable without flattening technical details.

• Write and maintain model cards and technical reports for new model releases.

• Work with researchers to help draft papers and reports for scientific audiences.

• Write practical guidance for integrating Pangram and interpreting results for customers, institutions, and developers.

• Standardize and maintain terminology, notation, citations, and formatting across documentation.

• Write and maintain user documentation covering product workflows, detection outputs, common questions, and recommended practices.

• Translate model behavior into documentation, including in-product guidance, FAQs, and explainers about prediction scores, accuracy, thresholds, uncertainty, and conditions that can affect reliability,

• Work with machine learning researchers to document model architecture, training methods, datasets, evaluation design, benchmark results, and known limitations in model cards and release notes.

• Draft and edit research reports and public-facing materials produced by the research team for clarity, consistency, and accuracy.

• Professional technical writing skills and experience.

• Strong structural and editing skills, with sound judgment about evidence, claims, and caveats.

• Degree in a quantitative field, such as statistics, computer science, data science, mathematics, or engineering, or related.

• Working familiarity with machine learning and model evaluation concepts, including classification metrics, benchmarks, and uncertainty.

• Ability to work independently with researchers and engineers to close gaps in source material.

• Comfortable creating order in a fast-moving environment; NYC-based and available to work with the team in person.

• Training in scientific and technical communication or computational linguistics or computational linguistics

• Familiarity with text classification, calibration, prevalence effects, and false-positive/false-negatives.

• Familiarity with text classification, calibration, and false-positive/false-negative tradeoffs.

• Experience in publishing, journalism, education, trust and safety, or other settings where technical outputs inform consequential decisions.

You can sit with a researcher, understand what changed in a model, find the central story in a dense set of results, and produce a draft that is technically faithful and comprehensive. You notice when a term shifts meaning, a number lacks context, or a conclusion outruns the data.

You are also highly self-directed. You create the outline, drive the review, keep the source of truth current, and know when to publish versus when to ask one more question. You care about language because it shapes how people use the technology, and you bring calm, sound judgment to writing about uncertain or high-stakes results.

• Salary: $100,000-$120,000 base salary.

• Equity: Meaningful ownership through stock options.

• Benefits: Health, dental, and vision coverage, plus a 401(k) with a 4% employer match.

• Time off: 20 days of PTO, company holidays, and a choose-your-own-holiday policy.