Nilesh Verma

—

Unclaimed · AI-structured profileOpenAlex

First author of TuiML, a self-describing machine-learning library that exposes ML models to LLM agents via MCP

papers here

reviewed by Evolve

first-author

corresponding

—

works · OpenAlex

—

h-index · OpenAlex

Bibliometric figures come from OpenAlex (2026-09-20) and cover the author's entire public record, not just the papers listed here. Reproducibility signal comes from Evolve's own reviews.

Research profile

Output—

Insufficient data: the indexed record contains only one reviewed output, so no publication-volume score can be computed. This axis would measure total output count, venue mix, and first/corresponding-author share across a fuller bibliographic record.

Impact—

Insufficient data: with a single reviewed paper and no citation statistics supplied, no impact score is available. This axis would measure citation uptake, influence beyond the immediate venue, and downstream reuse of the work.

Focus100/100

The focus score is 100: the indexed field distribution places all of this researcher's reviewed work — currently one paper — in a single field, machine learning for AI agent tooling. Every indexed output centers on that one topic, giving perfect topical concentration; note, however, that the score rests on a single indexed paper, so it reflects the record's narrowness rather than an established long-term research programme.

Collaboration28/100

The collaboration score is 28, based on 3 distinct co-authors in the indexed record. The frequent-coauthor list — Nick Lim, Albert Bifet, and Bernhard Pfahringer — suggests one small, stable working group, with the TuiML paper authored first-author alongside them. The score indicates a compact, tightly-knit collaboration footprint rather than a broad multi-team network, which is typical of a record resting on a single indexed paper.

Activity—

Insufficient data: the record's publishing cadence cannot be scored from a single 2026 preprint. This axis would measure publication frequency, year-over-year continuity, and recent output tempo, all of which need a longer publication history than the current indexed record provides.

Reproducibility55/100

The reproducibility score is 55, a mean taken over exactly one paper with medium repro-readiness. For TuiML, the Evolve review notes accuracy parity with scikit-learn on tabular benchmarks, which aids comparison, and reproducibility is among the paper's own keywords; yet the central claim that LLM agents benefit from the library was left untested, so the score sits mid-range pending stronger evaluation artifacts.

Research summary

Nilesh Verma is a researcher whose publicly indexed record — an early-stage profile resolved from bibliographic data — currently centers on one first-author paper: TuiML: Machine Learning for AI Agents (arXiv, 2026). The work presents a self-describing machine learning library designed for LLM agents, exposed through the Model Context Protocol (MCP) so that agents can discover and invoke ML functionality as a tool. Evolve's review characterizes the design as coherent and reports accuracy parity with scikit-learn on tabular benchmarks, while flagging that the headline claim — that agents themselves benefit from the library — remains untested; the paper carries medium repro-readiness and lists reproducibility among its own keywords. The bibliometric radar reflects how thin the record still is. Focus scores 100 because the field distribution places the single reviewed paper entirely within machine learning for AI agent tooling. Collaboration scores 28 from just 3 distinct co-authors, with Nick Lim, Albert Bifet, and Bernhard Pfahringer recurring — a compact working group rather than a wide network. Reproducibility scores 55 as the mean over that one paper. Output, impact, and activity lack sufficient data to score. No institution, country, or ORCID is available in the supplied data, so this profile should be read as an early-stage, name-resolved snapshot of a researcher just entering the indexed literature.

Collaboration pattern

The indexed record lists 3 distinct co-authors, and the recurring names — Nick Lim, Albert Bifet, and Bernhard Pfahringer — point to a single small, stable group: the TuiML paper was written first-author with these three. The collaboration score of 28 reflects a compact footprint rather than a broad network, consistent with an early record resting on one indexed paper. Broadening beyond this core team would materially change the collaboration picture.

Research keywords

AI agentsLLM tool useModel Context Protocol (MCP)Machine learning librariesSelf-describing library designScikit-learn compatibilityTabular machine learningTabular benchmarksReproducibilityAgentic workflow tooling

Highlights

TuiML: a self-describing ML library aimed at LLM agents
The 2026 first-author arXiv paper 'TuiML: Machine Learning for AI Agents' presents a machine learning library exposed to agents through the Model Context Protocol (MCP), so ML functionality becomes discoverable and invocable as a tool for LLM agents.
Accuracy parity with scikit-learn on tabular benchmarks
Evolve's review reports that TuiML reaches accuracy parity with scikit-learn on tabular benchmarks — evidence that the ML core performs like established tooling — while stressing that the claimed benefit on the agent side has not yet been tested.
Medium repro-readiness with the core claim untested
The paper carries medium repro-readiness, and the radar's reproducibility score of 55 is the mean over this single paper; reproducibility is among the paper's own keywords, but the Evolve verdict flags the agent-benefit claim itself as untested — the clearest next step for this line of work.

Timeline

2026Published first-author arXiv preprint 'TuiML: Machine Learning for AI Agents', introducing a self-describing ML library for LLM agents via MCP
2026Paper entered the Evolve catalog with a medium repro-readiness rating; the review cites coherent design, scikit-learn accuracy parity, and an untested agent-benefit claim
2026Indexed record shows 3 distinct co-authors, with Nick Lim, Albert Bifet, and Bernhard Pfahringer as the recurring collaborator group

Papers in the Evolve catalog

2026-09-17first author
TuiML: Machine Learning for AI Agents

Frequent co-authors

Nick Lim×1 Albert Bifet×1 Bernhard Pfahringer×1

About this profile

This profile is AI-structured by Evolve.Pub from public bibliographic data (OpenAlex) and Evolve's own paper reviews. It is not written or endorsed by the researcher unless the profile is claimed. Radar dimensions are computed from real bibliometrics with fixed formulas — raw figures are shown alongside; missing dimensions mean insufficient data, never zero. Errors? Claim the profile to correct or hide it.

Bibliometric scores derive from a limited indexed record; no citation counts or venue history were supplied.
The profile rests on a single reviewed paper (a 2026 arXiv preprint); output, impact, and activity axes lack sufficient data to score.
No institution, country, or ORCID was provided; the profile is name-resolved from public bibliographic data and may merge or split same-name authors.
The reproducibility score is the mean over one paper (reproMean 55, reproPapers 1), so it should be read as indicative rather than established.