Machine-learning researcher bridging low-resource Bangla medical NLP and uncertainty-aware solar PV forecasting
2
papers here
2
reviewed by Evolve
2
first-author
0
corresponding
—
works · OpenAlex
0
h-index · OpenAlex
Bibliometric figures come from OpenAlex (2026-10-02) 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—
Not yet scored: the snapshot lacks reliable output-volume data beyond one 2026-indexed work, so any estimate would be speculative. This axis would measure cumulative scholarly production — preprints, conference and journal papers — over the research career.
Impact—
Not yet scored: citation data is effectively empty (h-index 0, i10-index 0), and the two catalogued preprints are too recent for citations to accrue. This axis would measure citation-based influence over time, including citation counts and h-index trajectory.
Focus0/100
Scores 0, reflecting an even 0.5/0.5 field split between machine learning for renewable energy forecasting and clinical natural language processing — two communities with little overlap. Both 2026 preprints sit on opposite sides of that divide, so the record does not yet reveal a consolidating research direction. If later work converges on either low-resource NLP or energy forecasting, this measure should rise as topical concentration emerges.
Collaboration28/100
Scores 28, based on a thin collaboration footprint: three distinct coauthors in the record, with K. M. Tahlil Mahfuz Faruk and Md. Maruful Islam Maruf listed as the recurring collaborators. Both catalogued papers are first-authored without corresponding-author flagging, consistent with an early-career pattern of small-team work that has not yet branched into larger multi-institution consortia.
Activity20/100
Scores 20: the recent three-year window shows only one work, and the year-by-year record contains a single indexed output in 2026. With both catalogued preprints appearing in 2026, the visible activity is recent and clustered rather than sustained; a longer observation period is needed before a meaningful productivity trend can be established.
Reproducibility55/100
Scores 55, the mean across two reproducibility-assessed papers, both rated medium readiness. Evolve's verdicts flag concrete gaps: the Bangla NER benchmark admits its state-of-the-art margin reflects run variance, and the solar PV study is simulation-only and single-seed. Still, full-test-set evaluation and a reported coverage figure of 94.3% indicate disciplined reporting that could support replication with modest extra validation.
Research summary
Rakib Abdullah is an early-stage machine-learning researcher whose public bibliographic footprint, resolved via OpenAlex, currently consists of two 2026 arXiv preprints, both with the researcher in the first author position. The two works sit in deliberately different problem spaces: the field distribution is split evenly (0.5/0.5) between machine learning for renewable energy forecasting and clinical natural language processing. The first paper benchmarks transformer encoders for Bangla medical named entity recognition, reporting XLM-RoBERTa at F1 0.5959 ahead of BanglaBERT on a full-test-set evaluation, while candidly attributing the state-of-the-art margin to run-to-run variance. The second applies transfer learning with conformalized quantile regression to solar PV forecasting under load-shedding-driven data scarcity, cutting forecast RMSE by up to 23.7% and reaching 94.3% coverage in simulation, though it is simulation-only and single-seed. Quantitative impact indicators remain at zero (h-index 0, i10-index 0, one work in the recent three-year window, one indexed output in 2026), so this profile reads as an early snapshot of an active author rather than an established record. Evolve's reproducibility assessment averages 55 across two scored papers, both rated medium, suggesting reusable pipelines that still lack full validation. Collaboration is modest: three distinct coauthors, with K. M. Tahlil Mahfuz Faruk and Md. Maruful Islam Maruf recurring.
Collaboration pattern
The collaboration footprint is small: three distinct coauthors, and the recurring names in the record are K. M. Tahlil Mahfuz Faruk and Md. Maruful Islam Maruf. Both 2026 preprints are first-authored without a corresponding-author designation, fitting a small-team, early-career pattern. No multi-institution or large-consortium work is visible in this snapshot, so cross-group collaboration cannot yet be assessed.
Research keywords
Highlights
Bangla medical NER benchmark puts XLM-R ahead of BanglaBERT
Transfer learning plus conformalized quantile regression for PV forecasting under data scarcity
Dual-track first-author output spanning two unrelated ML domains
Both 2026 catalogued papers are first-authored yet sit in disjoint fields — clinical NLP and renewable energy forecasting — sharing a methodological thread of rigorous evaluation of pretrained and uncertainty-aware models. The field distribution records an even 0.5/0.5 split.
Timeline
- 2026Published a first-authored arXiv benchmark of transformer models for Bangla medical NER (clinical natural language processing).
- 2026Published a first-authored arXiv study applying transfer learning with conformalized quantile regression to solar PV forecasting under data scarcity.
- 2026OpenAlex year-by-year record shows one indexed work in 2026, with no earlier indexed output in the snapshot.
Papers in the Evolve catalog
Frequent co-authors
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.
- Bibliometrics derived from OpenAlex only; the h-index (0) and i10-index (0) reflect a sparse, early record rather than a mature citation history.
- Institution, country, and ORCID are absent from the data, so affiliation-level analysis and disambiguation were not possible; Rakib Abdullah may be a common name and identity is resolved from public bibliographic data.
- Only two papers, both 2026 arXiv preprints, are in scope; no peer-reviewed journal or conference publications are captured.
- The frequent-coauthor list includes the researcher's own name, suggesting imperfect name disambiguation in the underlying record.