We work as an embedded modelling team for biotech and pharma clinical pharmacology groups. Each engagement is a defined package of datasets, model, code and report, priced and timelined before kickoff.
Structural and stochastic model development for SAD/MAD, dose-finding studies, and pivotal trials. Full covariate analysis (forward inclusion / backward elimination or full random-effects modelling), bootstrap, VPC, and prediction-corrected diagnostics. We are comfortable working in NONMEM, Monolix, nlmixr2 and Pumas.
Continuous biomarkers, categorical endpoints, count data, and time-to-event analyses. Indirect-response models for delayed PD, logistic regression for binary endpoints, Cox models with time-varying exposure metrics for efficacy and safety endpoints. Output: defensible dose-justification narratives.
Clinical trial simulation to inform sample size, dose selection, sampling design, and adaptive design rules. Pediatric extrapolation under ICH E11A. Bridging analyses across populations and formulations. We work directly with development teams to make MIDD a decision tool, not a deliverable.
End-to-end pharmacometric write-ups for CTD Module 5.3.3 (PopPK reports) and 5.3.4 (exposure-response). Reviewer-ready analysis datasets and code archives. Pre-IND and pre-NDA briefing-document support. We have submitted to FDA, EMA, PMDA, Health Canada and MHRA, and we write to be defensible at review.
Standalone C-QTc analyses to support TQT waivers per ICH E14 Q&A (R3). Pre-specified analysis plans, linear mixed-effects models with heart-rate correction, sensitivity analyses, supratherapeutic scenario predictions, and the supporting CSR sections.
DDI prediction (perpetrator and victim), organ-impairment scenarios, pediatric scaling, and food-effect simulations. Built in PK-Sim / MoBi with verification against observed clinical data and literature. We follow the FDA/EMA PBPK qualification expectations and document accordingly.
For teams that need ongoing pharmacometrics capacity without hiring. We embed a senior pharmacometrician with your clinical pharmacology group on a rolling quarterly basis, with our agent stack behind them. Predictable monthly fee, defined deliverable cadence.
Four phases. Each one ends with a written artefact that your team can review, share with regulators, or hand to the next vendor without us in the loop. From the day your data lands to a submission-ready report, the whole sequence runs in six weeks or less.
30-minute call to review the question, the dataset, and the regulatory context. One-page scope document and fixed-price proposal within five working days. If we are not the right fit, we say so.
Our internal agents assemble analysis-ready datasets from raw clinical data: CDISC-aligned, fully traceable, versioned in git. A senior pharmacometrician reviews and signs off before any modelling.
Structural and stochastic model development, covariate analysis, diagnostics, simulation. Agents handle code drafting and run-record bookkeeping; modelling decisions stay with the scientists. Weekly written updates.
Final report, analysis datasets, NONMEM/Monolix archive, code repository, and a working session with your team. Every artefact is built to be lifted straight into your CTD.
Our agents are a set of proprietary tools, each one specialised in a single step of the pharmacometrics workflow, built because off-the-shelf options were not good enough for regulatory-grade work. They run inside our environment, under our review, and they are why a full analysis takes weeks rather than months. They do not make modelling decisions and they do not write the report.
Pulls from raw SDTM, harmonises units, flags BLQ records, builds the NONMEM-ready dataset with full traceability back to source.
Drafts NONMEM control streams, R analysis scripts, and diagnostic plotting code from a model specification. A scientist edits and runs them.
Captures every run, parameter, diagnostic, and decision. Flags numerical instabilities. Produces the audit trail expected by reviewers.
Our agents are not a thin layer over a public AI product. They run on infrastructure we own and control, inside a closed environment your trial data enters under NDA and never leaves. Where the work runs, and what happens to your data, is agreed in writing before the project starts.
The agents are not a fixed, off-the-shelf pipeline. For every engagement they are configured to the data, the modelling plan and your internal SOPs, then run as one orchestrated team under a named pharmacometrician who directs them and signs off each step.
Everything runs on private, access-controlled servers we own and operate, not on public cloud AI services. Your trial data is never sent to a third-party model or an external API.
Client data is brought into the environment under NDA, stays inside it for the whole project, and is never used to train, fine-tune or improve any model, whether ours or anyone else's.
Send a protocol synopsis or a short description of the problem. We come back within five working days with a written scope and a fixed price.