ICLR 2026 LLM Reasoning Workshop
I make complex things understandable: in AI research, live talks, and online courses.
I'm Olga Manakina, a cognitive scientist specializing in applied AI and human learning. My research examines how LLM agents behave over time and how AI can be used to enhance learning and streamline teaching practices. Over 2025–26 I published and presented my work at ICLR, NeurIPS, AIED, EDM, and IEEE COMPSAC. Before my transition to AI research, I spent years designing and delivering courses for university students and adult learners. My online products grew organically to 35,000+ subscribers, whose most common reaction to my courses is: after your explanation it finally clicked.
Questions I work on
How do LLM systems behave over long interactions?
Reliability, temporal failure patterns, reasoning consistency, and agent evaluation.
When does AI support learning, and when does it replace the thinking that produces learning?
Engagement, scaffolding, mastery learning, and human–AI interaction.
Can LLMs coordinate complex processes reliably, rather than perform isolated tasks?
Instructional orchestration, state tracking, grading, progression, and remediation.
How can AI systems become more efficient without sacrificing quality?
Adaptive prompt selection, evaluation design, and accuracy–cost trade-offs.
By the numbers
5peer-reviewed
papers
2025–26
papers
2025–26
35.2KYouTube
channel
subscribers
channel
subscribers
1,030+Udemy
course
learners
course
learners
Research toolkit
Study DesignExperimental Design · Mixed-Methods Research · Benchmark Design · Human–AI Studies
AnalysisStatistical Analysis · Survival & Hazard Modeling · Qualitative Coding · Taxonomy Development · Inter-Rater Reliability
AI Systems & EvaluationLLM Evaluation · Agentic Workflows · Prompt & System Design · LLM-Assisted Data Extraction · Python · Pandas
Research overview
Research Highlights
Five peer-reviewed studies and my MA thesis. Each has an interactive explainer and a link to the publication.
AIED 2025 Main track
Three simple engagement prompts more than doubled mean learning gains.
NeurIPS 2025 MTI-LLM Workshop
Persona prompts strongly changed failure risk—and mandatory self-questioning increased it.
EDM 2025 WLIEDM Workshop
A bandit learns which grading strategy to favor while it grades—using 78.4% fewer LLM calls than exhaustive search.
IEEE COMPSAC 2025 QUORS Workshop
A framework for integrating AI into software engineering curricula, grounded in learning theory and expert-reviewed before student testing.
MA thesis 2026 Carleton University
A prompt-configured LLM coordinated a complete mastery-learning cycle with high, but imperfect, reliability.
Digital portfolio
Educational Products
YouTube · 2020–present
Educational technology channel
54 instructional videos teaching educators to run online classes (Zoom, Loom, tablets as whiteboards, Google Slides), grown organically to 35.2K subscribers and 2.9 million views through audience-informed content refinement. No ads, no cross-promotion.
35.2K subscribers · 54 videos · 2.9M views
Udemy instructorLearn French Onlinewith Dr. Olga Manakina
Udemy · 2018–present
"Learn French Online" course series
Five French-language online courses, from phonetics for absolute beginners to DELF B2 exam grammar, authored, recorded, and maintained end-to-end, serving over 1,030 students and refined using learner needs analysis and engagement data.
5 courses · 1,030+ learners · 369 reviews · 4.7★ average
Recognition
Awards
Russia · 2015
Government of the Russian Federation Prize in Education
Co-author of the awarded course package Le Français.ru B1.
Russia · 2010
National Prize "Book of the Year"
Nomination "Textbook of the 21st Century". Co-author of the awarded course package Le Français.ru, levels A1–B1.