ML researcher spanning interpretable LNP drug-delivery prediction and efficient long-context LLM inference
2
papers here
2
reviewed by Evolve
0
first-author
0
corresponding
—
works · OpenAlex
—
h-index · OpenAlex
Bibliometric figures come from OpenAlex (2026-09-19) 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—
Publication-volume data for this axis is not yet available in the input, so no score is shown. Output would measure the total number of cataloged papers and preprints over time, beyond the two 2026 arXiv entries reviewed here.
Impact—
No citation or influence metrics were provided, so impact cannot be scored. This axis would measure citations, citation performance relative to field, and uptake of the two reviewed preprints, both recent 2026 arXiv postings without recorded citation data.
Focus0/100
The focus score is 0, driven by an even 0.5/0.5 field distribution between computational nanomedicine and machine learning systems/LLM inference across the two cataloged papers. One paper predicts LNP biodistribution with interpretable ML; the other targets speculative decoding for long-context inference. The score reflects a deliberately two-sided portfolio, though it is computed from only two reviewed papers and may shift as more work is cataloged.
Collaboration48/100
The collaboration score is 48 (moderate), based on 10 distinct coauthors in the cataloged network. Frequent coauthors include Asal Mehradfar, Mohammad Shahab Sepehri, Owen Antholine, Varun Shankar, Glen S. Kwon, Morteza Rasoulianboroujeni, Amir Ziashahabi, Hossein Entezari Zarch, Lei Gao, and Murali Annavaram, a mix consistent with the two-field split. Neither reviewed paper lists him as corresponding author, with author positions of middle and last.
Activity—
Activity data such as publication cadence and yearly output is not available in the input, so this axis is unscored. It would measure how regularly new papers appear over recent years, which cannot be inferred from the two 2026 arXiv preprints alone.
Reproducibility100/100
Reproducibility is the strongest measured axis: a mean score of 100 across the 2 reviewed papers, both rated high reproducibility readiness. The LNP machine-learning study builds on a curated 476-formulation dataset with interpretable SHAP analysis, and ASPIRE reports concrete 1.70–4.58× speedups. Review caveats — possible study-level split leakage and thin variance reporting — temper interpretation but do not lower the readiness ratings.
Research summary
Salman Avestimehr is a machine-learning researcher whose cataloged work spans two distinct areas: computational nanomedicine and efficient LLM inference. This profile is name-resolved from public bibliographic data; no ORCID or institutional affiliation was provided in the input. Evolve's catalog contains two 2026 arXiv preprints. In computational nanomedicine, he is a middle author on work applying interpretable machine learning to lipid nanoparticle targeting: models trained on 476 curated LNP formulations predict liver versus extrahepatic IVIS outcomes with AUC up to 0.874, using SHAP analysis and RDKit molecular descriptors to extract formulation design rules. In ML systems, he is the last author on ASPIRE, an asynchronous batched self-speculative decoding method for long-context LLM inference that de-synchronizes speculative execution through mixed forwards and per-request scheduling, with reported speedups of 1.70–4.58×. His cataloged field distribution splits evenly (0.5/0.5) between the two areas, and the focus score of 0 reflects that balance rather than narrow specialization. He collaborates widely relative to catalog size — 10 distinct coauthors — with a moderate collaboration score of 48, and serves as neither first nor corresponding author on either reviewed paper. Both papers carry high reproducibility readiness and a perfect reproducibility mean of 100, though review verdicts flag possible study-level split leakage in the nanomedicine work and thin variance reporting in ASPIRE.
Collaboration pattern
Salman Avestimehr's cataloged network contains 10 distinct coauthors, giving a moderate collaboration score of 48. Frequent coauthors include Asal Mehradfar, Mohammad Shahab Sepehri, Owen Antholine, Varun Shankar, Glen S. Kwon, Morteza Rasoulianboroujeni, Amir Ziashahabi, Hossein Entezari Zarch, Lei Gao, and Murali Annavaram. His roles are collaborative rather than corresponding-author led: middle author on the nanomedicine paper and last author on ASPIRE, with the network spanning both computational nanomedicine and ML systems.
Research keywords
Highlights
Interpretable ML decodes extrahepatic targeting of lipid nanoparticles
ASPIRE: asynchronous self-speculative decoding for long-context LLM serving
Timeline
- 2026Co-authored the interpretable machine-learning study decoding extrahepatic targeting of lipid nanoparticles (middle author, arXiv).
- 2026Co-authored ASPIRE, asynchronous batched self-speculative decoding for long-context LLM inference (last author, arXiv).
- 2026Both papers entered the Evolve catalog with high reproducibility readiness and a 100 mean reproducibility score; the cataloged network spans 10 distinct coauthors across the two fields.
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.
- Radar bibliometrics are derived from OpenAlex-based data only; the output, impact, and activity axes lacked sufficient data to score.
- Profile is name-resolved from public bibliographic data; no ORCID, institution, or country was provided, so identity confidence is limited.
- Scores rest on only two Evolve-reviewed 2026 arXiv preprints without stated peer-review status, so focus and collaboration values may change as the catalog grows.