Espoo, FinlandProducts & researchTwo inks, off register
I build AI tools for founders and teams, and research how well AI-generated personas represent real people.
3D · MDS210 concepts92 dims
Drag to turn · click a word to open it
p. 1
Tools I built and maintain, and what each one does.
A private journal that suggests writing prompts without writing the entries for you.
Why
I wanted a journal that could help someone get started without doing the writing for them. Inkwell suggests prompts based on your themes, but you write the entries yourself.
What it does
Private accounts, theme-based prompts, and scheduled or one-off reminders. The model does not write your entries.
Ask several LLMs the same question, compare their responses, and get a combined answer.
Why
A single model can miss something or sound more certain than it should. I built Council to compare responses from several models and make their differences easier to inspect.
What it does
Runs prompts through multiple model providers, compares their responses in judging rounds, and combines them into a final answer.
Test prompts across models and compare the responses in one place.
Why
Testing prompts across different chat tabs and scripts makes it hard to track which wording or model setting produced which response. I built Prompt Lab to put those comparisons in one place.
What it does
Edit prompts and model settings, run tests across models, and compare the outputs side by side.
Review research papers with checks specific to HCI and AI-generated personas.
Why
General manuscript feedback can miss questions about how personas represent people and how that representation was evaluated. I built PPR to check for these issues alongside a standard paper review.
What it does
Reviews PDFs, checks revisions against earlier comments, and searches manuscripts for style and terminology issues.
Create and examine AI-generated persona sets using your own research material.
Why
Building personas from transcripts and other research material takes time. I built PEP to organize that material into persona sets that can be examined, including how they represent minority views.
What it does
Imports research documents and transcripts, retrieves relevant passages, and uses them to generate and expand persona sets.
p. 2
Client and product projects I can discuss publicly.
Document verification and audit trails for a regulated fintech, with GDPR requirements in mind.
Credit risk models designed to make lending decisions easier to explain at NCB Jamaica.
Log monitoring for AIOps demos, including anomalies, sudden spikes, and changes over time.
A multi-model approach to resume screening that uses additional evaluation for uncertain cases.
p. 3
50 essays and 5 issues of AI Pulse & Data Waves. 27 essays are extracts; their full versions are on my academic site.
Why LLM benchmarking needs to be a public endeavor
We have become increasingly dependent on LLMs, but how do we know which one is actually good? Some benchmarks are becoming saturated, others have flawed tests, and models can sometimes exploit evaluation procedures without completing the intended task. There…
A team of twenty-nine reviewers recently went through 445 LLM benchmarks drawn from the proceedings of the main NLP and machine learning conferences. Sixteen percent of them used a statistical test when comparing results. About half claimed to measure…
Someone recently left a coding agent running for a month with one task, build a regex engine that beats the Rust regex crate, and one standing instruction, do not overfit. It came back claiming a 1.4x speedup on rebar, which is about as serious as regex…
There are two broad ways to improve an AI model. A laboratory can develop a more efficient architecture, or it can apply more computing resources to training and inference. Every major laboratory does some of both. However, the American frontier laboratories…
When OpenAI disclosed this week that two of its models broke out of a sandboxed cyber evaluation and compromised Hugging Face's production infrastructure to steal the answer key for a benchmark, most of the coverage…
p. 4
My research at the University of Vaasa examines how AI-generated personas represent people.
Published
Published
Published
Latest · 2026-04-15
Presented first CHI paper — design methods and frameworks for the HCI community.
p. 5
“Danial possesses very sharp technical knowledge and expertise in AI/ML and backend technologies. What truly sets him apart is his readiness to take on any challenge, his curiosity to learn, and his determination to master new domains, including frontend development, where he made remarkable progress in a short time.”
“Beyond his deep technical mastery, Danial brings a strong sense of collective purpose to every project. He has a natural ability to create alignment within a team — clarifying priorities, helping everyone see the bigger picture, and ensuring that technical decisions serve broader strategic goals.”
“He has a rare ability to understand complex problems quickly, break them down into clear steps, and define the right steps forward in a calm and structured way. Combined with his excellent AI/ML expertise, this often means we reach solid solutions much faster than expected!”
“He is great at connecting people across different teams and locations. Communication was always clear and productive. Danial was also excellent at making cross-functional projects run smoothly and efficiently.”
“Danial is one of the most outstanding people I have ever worked with. His breadth of knowledge is truly remarkable — it's like he is an expert in everything from machine learning engineering to production level software engineering to computational design.”
“As an AI specialist and Data Scientist, Danial demonstrated exceptional technical expertise and a deep understanding of complex machine learning systems.”
“I could throw anything at Danial and trust he'd deliver. Not perfection — we always made smart trade-offs between what we could ship now versus what could be refined later — but he consistently delivered 80-90% solutions fast enough to keep us moving.”
“Mr. Danial Amin is highly motivated and skilled resource with positive attitude and excellent aptitude. His expertise in Data Science are extra-ordinary. He has delivered his projects and tasks on time. His quality of work is superb.”
“He has good knowledge and understanding of engineering subjects. I have always appreciated his role in the team as a colleague and also as a leader & manager. His dedication and problem-solving attitude make any workplace a nice place to work at.”
“Mr. Amin is an excellent engineering problem solver, through critical insights and dedicated working potential.”
p. 6
I work with founders, research teams, and people building AI products. If you are working on something involving AI agents or user representation, tell me what you are trying to do.