Bhanoji Duppada · Principal Statistical Programmer & Platform Developer
Clinical programming systems you can put in a submission.
I design, build and run the tools that take a study from raw data to an FDA package: SDTM, ADaM, TLF, define.xml. Everything below is live behind HTTPS on infrastructure I operate, and every number on this page is counted, not claimed.
Run R in the browser nowSee the systems
No login for the demo. Synthetic data only.
Track record
Leadership and delivery
11+ years in clinical statistical programming across Phase I to III oncology and neuroscience trials: leading teams and CRO delivery, owning independent validation, and shipping end-to-end FDA submissions. The platform above is that same discipline, automated.
Led teams and CRO delivery
As Senior and Principal Programmer at CRISPR Therapeutics, led programming personnel across oncology studies and managed CRO interactions, owning on-time delivery of analysis datasets and TLFs.
- Programming oversight and review of juniors and CRO deliverables
- On-time delivery under SOP, ICH and GCP
- Risk-based validation across concurrent studies
Independent validation
Drove independent double-programming of critical ADaM datasets and TLFs across oncology and neuroscience programs, reconciling discrepancies directly with Biostatistics and external partners.
- Parallel programming of ADSL, ADAE, ADTTE, ADRS
- Survival and efficacy endpoints: PFS, OS, TTP, DoR
- Pinnacle 21 conformance and discrepancy reconciliation
End-to-end FDA submissions
Delivered submission-grade packages end to end: SDTM and ADaM, define.xml, SDRG and ADRG, and .xpt transport datasets for eCTD and NDA, and contributed to database-lock activities.
- define.xml 2.x, cSDRG, ADRG, .xpt for eCTD
- ISS and ISE, pooled SDTM, CSR section packing
- Phase I to III, oncology and neuroscience
Platform architect
Channeled that submission discipline into clincoder.cloud: a suite of production clinical-programming tools from raw data to submission package, architected and operated end to end by one programmer.
- Paired SAS and R from one shell; cell-by-cell RTF parity
- CDISC ARS conformance and a synthetic ADaM generator
- Live, monitored services, not slideware
Systems
Raw data to submission package, in the order it happens.
Raw data → SDTM
SDTM Mapper live
Raw CSV or SAS7BDAT in; domain classified, a variable-by-variable mapping spec, SAS, R and Python programs generated in parallel, executed server-side, SDTM CSV and XPT out. Refusals are shown, not hidden.
SDTM Automation Agent live
The agent behind the mapper: raw data plus specs to SDTM datasets and SAS programs, metadata-aware and CT-validated, with an MCP server for seven tools.
SDTM → ADaM
ADaM → TLF
ShellForge live
Upload any table shell (DOCX, XLSX, RTF, PDF, TXT). Every row is annotated to ADaM dataset, variable, filter and statistic by a five-pass rule engine with a constrained LLM fallback, then paired SAS and R programs and an annotated PDF are generated.
ClinAssist live
A TLF shell specification to submission-style SAS and R code through a retrieval engine. Browser-side sanitisation keeps compound names away from the model; output carries a GAMP 5 audit header.
ARS, the machine-readable spec
ARDflow live
A CDISC Analysis Results Standard spec in; the Analysis Results Dataset computed and rendered as submission RTF, Dataset-JSON and HTML with a full spec-to-result trace. Pure R.
ARS Validator live
Paste or upload ARS JSON; a 21-rule conformance engine checks structure, references and controlled terminology and returns severity-graded findings.
TLF Viewer live
Import an ARS spec and see the table or listing shell it defines, before any SAS runs.
Submission package
QC and review
QC Cockpit live
A dashboard for TLF QC: production against QC, cell by cell, with the discrepancies listed rather than a green tick.
ClinSafe Monitor live
Pharmacovigilance signal detection over study data, on synthetic data here.
Learn and assist
RLab live
Learn clinical R by running it in the browser: 23 gated topics on synthetic ADaM, no login.
CVRM Figure Playground live
Sixteen CVRM efficacy figures — Kaplan-Meier, forest, MCF, win ratio, waterfall, competing-risk CIF, eGFR slope — each with the PROC SGPLOT option playground behind it, the generated SAS and R, and the table-to-figure trace. Synthetic data, no login.
Ask a Clinical SAS Programmer live
Free 15-minute expert sessions on CDISC, SDTM, ADaM, TLF and define.xml questions, grounded in a curated standards knowledge base.
AI governance & validation framework live
How model-assisted programming is validated and governed under GxP: roles, evidence and the gates a generated program must pass.
Only systems that answered are shown as live. A system that goes down is listed, not quietly removed.
Method
Only a run writes evidence.
Every gate has to fail first
A check that has never failed has never been shown to work. Each system ships with a deliberate-failure leg that fired before it went live.
N of M, never a percentage
One run is an observation, not a rate. Counts keep the denominator visible; a bare percentage hides how much a single run can move.
Synthetic data, said out loud
Every dataset here is synthetic or a public standard. No sponsor, compound or study is named anywhere on these pages.
Writing
The notes behind the systems.
Field notes, decision tables and SOPs on the ClinCoder blog. Articles are never behind a login; every figure carries its run.
The PFS censoring decision table you can paste into a SAP
Progression-free survival censoring rules, written as one table you can drop into the SAP, define.xml and ADRG — plus the ADaM traps that survive review.
Risk scoring is not testing — a vendor-neutral SOP for R packages in GxP
Most R validation packages score risk and stop. Scoring tells you how much testing to do; it is not the testing. Here is the SOP that separates them.
Books and guides
Practical references for clinical programmers.
- Clinical SAS Programming: The Complete Guide
- SAS to R: A Clinical Programmer's Guide
- SAS to R, Interactive Edition
- 100 SAS Macros for Clinical Programming
- Pharmaverse in Practice
- SAS Clinical Interview Cracker: 200+ Q&A
- TLF Automation Cookbook: SAS & R
- CDISC Decoded: A Practical Guide
- AI-Powered Clinical Programming
- Codebase Study Packs
Contact
Talk to me about your submission pipeline.
Based in the United States, working remotely. Teams, recruiters and collaborators welcome; a reply within one business day.