FinTech
Understand transaction, settlement and reconciliation exports, flag unmatched records and retain the transformation record.
Not a certified reconciliation engine.What is Data Sculptor? A local-first data engineering application under development, designed to make difficult data preparation reproducible, understandable, and accountable — on computers ordinary people already own.
A recurring spreadsheet, a financial export, a research dataset or a creative archive may outlive the person who created it. The result survives, while the instructions, decisions, exclusions and evidence disappear across different tools.
Data Sculptor is designed to keep the work behind the answer: what was requested, what ran, what changed, what was refused and which checks support the output.
State the inputs, transformations and rules explicitly.
Run supported operations; refuse ambiguity under defined rules.
Record what the execution actually did and reported.
Check supported properties and preserve bounded evidence.
Carry outputs, instructions and evidence in a portable package.
The architecture also proposes permanent row identifiers (RIDs) for record traceability. A Receipt is evidence of execution — not a blanket promise of correctness or a certified audit trail.
Planned Rust and WebAssembly processing in a browser-based Workbench, local-first data handling, streaming and disk-backed techniques, plus portable UTF-8 CSV and Excel exports.
Supported analytical operations are planned through readable DS-STEPS, DS-SQL and DS-Pipe forms that share the same underlying semantics.
Prepare data and preserve the transformation record, then hand off results to tools such as Tableau and Power BI.
Those established products already offer preparation workflows. Data Sculptor's proposed differentiator is the combined evidence model and portable local-first experience, which still requires comparison and validation.
These are proposed users and testable use cases, not seven separate shipped industry editions.
Understand transaction, settlement and reconciliation exports, flag unmatched records and retain the transformation record.
Not a certified reconciliation engine.Prepare matter inventories and structured evidence metadata with explicit exclusion and processing decisions.
No e-discovery or admissibility certification.Prepare authorized research and operational datasets through reproducible, local-first workflows.
Not a clinical device or certified health-data environment.Turn repeated CSV and Excel chores into clearer, repeatable Work Orders with portable outputs.
Supported operations will expand with implementation and testing.Explore asset inventories and metadata for photographers, musicians, filmmakers, journalists, authors, editors and publishers.
Full digital asset management is a future research direction.Make donor, alumni, grant and program exports easier to prepare and account for on limited budgets.
Privacy safeguards require specific review and testing.Explore transparent preparation of statistical, program and operational datasets on approved local infrastructure.
No procurement, security or accessibility certification is claimed.Digital asset management (DAM) is a researched future direction. For images, music, film and writing, the aim is to help organize original files, versions, descriptive metadata, credits, rights and related evidence. That might support RAW/JPEG pairs, audio mixes and stems, video clips, interview notes, manuscripts, citations and publication packages.
Data Sculptor is not currently presented as a complete DAM system, word processor, citation manager or media production suite.
Paired Long-Division Development (PLDD) separates requirements, adversarial specification review, frozen decisions, RED-to-GREEN checkpoint implementation, independent challenge and human acceptance. Founder and Product Manager Tarjin Rahman remains responsible for decisions and acceptance; a specialized AI Council supports architecture, implementation, QA challenge and communication.
Cost and schedule may change what we build and when we release it. They may never change what PASS means.
As of Day074, 15 requirements are partially touched, 387 have not started, and zero requirements have been fully accepted. Rustfmt and a WASM library build passed at CP019. These results do not establish browser runtime completion or production readiness.
Day075 (22 October): research and product vision, zero code checkpoints.
Day076 (23 October): website and evidence-maintenance preparation, zero code checkpoints.
No claim is made of financial, legal, healthcare or government certification, completed creative DAM, verified production throughput, universal security or product-market fit.
Read the public WASM-002 Development Log ↗ Founder’s professional portfolio ↗