Ahsan Habib ShihabResearch & practice

LLM EVALUATION / MULTILINGUAL NLP / AI RELIABILITY

Almaty, Kazakhstan / 2026

DATA SCIENCE & AI RESEARCH

Ahsan Habib
Shihab.

Portrait of Ahsan Habib Shihab

I study the reliability of AI evaluations, with a focus on cross-lingual consistency, trustworthy language models, and evidence-driven assessment.

Final-year Data Science undergraduate & Research Programmer at Al-Farabi Kazakh National University. Working across multilingual NLP, agent governance, and safety-sensitive applications.

Explore the research About me

Seeking MSc/PhD research opportunities for Fall 2027.

Halyk Academy
National Kazakh-language LLM initiative · BR24993001

IN PROGRESS

Evaluating historical knowledge through primary texts and expert review.

Al-Farabi AI Avatar

01SELECTED WORK

Research questions.
Practical responsibilities.

All projects

01SANA / OFFICIAL STATISTICS

Official statistics, in everyday language.

Making official statistical microdata accessible through natural-language questions.

Explore Sana

CONTRIBUTION & CONTEXT

My role
Core AI system lead · research methodology · evaluation design · technical coordination
Recognition
3rd place, AI for Official Statistics track, STAT.DATATHON-2026 — Team Sana, October 2026.
Evidence
Team-reported internal evaluation. Scope and limits in the project dossier.

02 / SOURCE-GROUNDED GENERATION

Al-Farabi AI Avatar

Evaluating historical knowledge with primary texts and interdisciplinary expert review.

Primary textsEvaluation corpusExpert review
Ongoing research

02PUBLISHED RESEARCH

The work, on record.

Full research record

Accepted work and conference presentations are identified separately in the research record.

RESEARCH MATERIALS

Code & Reproducibility

Explore code & artifacts

03A WORKING PERSPECTIVE

Language.
Evidence.
Consequences.

What a model scores
is only part of the story.

My research asks what makes an AI system reliable when language, context, or consequences change. That connects my work on Kazakh-language models, agent authorization, and safety-sensitive applications.

Research interests Teaching, service & background

OUTSIDE A SINGLE PAPER

Building, teaching,
and continuing the question.

Orbitum began with a hackathon we didn’t win. Continuing with a new team became sustained work. Alongside research, I teach multi-agent workflows and explore the connections between language, education, and climate resilience.

More about my path

↗CONTACT & COLLABORATION

Have a question
in common?

I welcome conversations about research, graduate study, and collaborations in AI evaluation, multilingual NLP, and reliable AI systems.