Alec Glisman, Ph.D.

Alec Glisman, Ph.D.

Senior AI/ML Scientist at Merck & Co. Ph.D., Chemical Engineering, Caltech 2024.

Product

Product

Platform

web

Users

Primary: industry recruiters and hiring managers evaluating Alec for senior AI/ML scientist roles in drug discovery / computational chemistry, deciding whether to reach out or advance him in a process.

Secondary (roughly equal weight, but the site should lean toward the recruiter read first): academic and industry research collaborators, conference attendees, and prospective co-authors assessing his publication record and research direction.

Product Purpose

A personal academic/professional site that serves as the canonical hub for Alec Glisman’s research, publications, experience, and skills. Success is a visitor (recruiter or collaborator) quickly understanding his expertise and credibility, then taking action: reaching out, advancing a hiring process, or pursuing a collaboration/citation.

Positioning

Sits at the intersection of physics-based simulation, machine learning, and molecular design — grounded in molecular-scale physics (soft matter, polyelectrolyte complexation, membrane mechanics) and applied to generative AI and property prediction for drug discovery. The combination of deep physical-simulation training with production generative-AI/ADMET work for a pharma company is the differentiator most neighboring ML-only or chemistry-only candidates can’t claim.

Operating Context

  • Ph.D. Chemical Engineering, Caltech (2024); MS Caltech (2022); BS UC Berkeley (2019).
  • Currently Senior AI/ML Scientist at Merck & Co., working on generative models (GFlowNet-based R-group optimization, core-hopping) and ADMET property prediction pipelines deployed to medicinal chemistry teams.
  • Prior doctoral research: molecular dynamics (metadynamics, replica exchange) of polyelectrolyte-ion binding in collaboration with Dow Chemical; earlier work in microhydrodynamics/active matter and lipid membrane continuum mechanics.
  • Site sections: Research, Publications, Experience, Projects, Skills, Talks, Teaching (see _data/navigation.yml).

Capabilities and Constraints

  • Static Jekyll site on GitHub Pages (Academic Pages / Minimal Mistakes fork); content lives in _pages/, _publications/, _talks/, _teaching/ collections.
  • No backend, forms, or dynamic data — all content is static Markdown/YAML, rebuilt on push to main.
  • Contact is via external links (LinkedIn) rather than an on-site form.

Brand Commitments

  • Name: Alec Glisman, Ph.D.
  • Voice: precise, physics/ML-literate, first-person, technically substantive rather than promotional (e.g. “I want to understand why molecules behave as they do”).
  • Visual identity already established: Navy (#1B2A4A) + Teal (#0EA5C9) palette, Syne/Outfit/Inter type system, domain badges (AI/ML, Physics, Drug Design, Simulation) — see _sass/_variables.scss and _sass/_custom.scss.
  • Structural direction (confirmed 2026-09-13): conventional/category-standard page composition (identity hero → proof strip → content), not an experimental structure (split-screen sidebar, timeline-led, or research-grid-first were considered and declined). Craft bar set against two named references, both explicitly rejected as-is: meredithschmehl.com (too plain, lacks visual cues) and bjfogg.com (missing visual elements/cues). The standard applies to future page rebuilds unless the user changes it.

Evidence on Hand

  • Five real publications with DOIs, venues, and citations in _publications/ (Langmuir 2025/2024, Macromolecules 2024, J. Fluid Mechanics 2022, Physical Review E 2020).
  • Real quantified achievement: 25th of 100 finalists, OpenADMET-ExpansionRx Blind Challenge, with public leaderboard and code links.
  • Real code repositories (DDPM-Enhanced-Sampling, OpenADMET-ExpansionRx-Blind-Challenge) and figure assets under images/.
  • No testimonials, case studies, or press exist and none should be fabricated.

Product Principles

  1. Lead with the recruiter/hiring-manager read (credibility, current role, quantified impact) while keeping the research depth that academic visitors need one click away.
  2. Every claim must be traceable to real, verifiable evidence (publications, DOIs, code, competition results) — never invented metrics or endorsements.
  3. Voice stays substantive and physics/ML-literate, not marketing-toned; the differentiation is real technical range (physics simulation → generative AI), not styling.
  4. Content changes must keep the static-Jekyll, no-backend model — no feature should require a server, database, or form backend.