RSI TECH
Contact

Work / Engineering evidence

Systems explained by outcomes, constraints and architecture.

Each report shows the decision path behind delivered programs, active work, research labs and open-source systems. Generalized client context stays explicit; evidence comes before stack labels.

All reports

08 systems
active / LLM Systems / 2025 -> now

Career Survey Intelligence & Compensation Consistency Engine

Structured extraction and grounded consistency checks made compensation and leveling review explainable.

Constraint Accuracy and auditability matter more than cleverness.

Select a signal node to inspect its operating principle.

  1. ExtractionNormalize free text into structured signals.
  2. GroundingCombine models, rules, and domain constraints.
  3. ConsistencySurface leveling and compensation conflicts.
  4. Human reviewAttach rationales and traceable evidence.
LLM pipeline, consistency detection, optimization, and evaluation engineeringPython / LLM pipelines / Optimization
delivered / Data Platforms / 2022 -> 2025

Marketing Analytics & Incremental Lift Optimization

Incrementality-focused measurement and targeting models connected model outputs to budget and campaign decisions.

Constraint Incrementality must anchor optimization decisions.

Select a signal node to inspect its operating principle.

  1. MeasurementAnchor decisions in incrementality.
  2. TargetingModel CLTV, propensity, and look-alikes.
  3. RecommendationsConnect outputs to expected lift.
  4. MonitoringGuard against silent degradation.
Analytics frameworks, targeting models, reporting, and production integrationPython / Predictive modeling / Targeting
delivered / Security Analytics / 2019 -> 2022

Security Ops: User Behavior Analytics & Anomaly Detection

Behavioral baselines and explainable anomaly signals for SOC triage and investigation.

Constraint Data volume is high; latency and cost constraints matter.

Select a signal node to inspect its operating principle.

  1. TelemetryNormalize multi-source logs.
  2. BaselinesModel user, session, and device behavior.
  3. SignalsScore anomalies with controllable thresholds.
  4. SOC triageKeep every alert explainable.
Data pipeline, feature engineering, anomaly detection, and methodology documentationPySpark / Python / Unsupervised learning
delivered / Data Platforms / 2021 -> 2022

Automated Time-Series Forecasting Framework

A repeatable forecasting workflow made model comparisons safer and iteration easier to hand over.

Constraint Data quality, objectives, and evaluation discipline are part of the forecasting process.

Select a signal node to inspect its operating principle.

  1. DatasetPrepare consistent model inputs.
  2. BaselinesAnchor every comparison.
  3. CandidatesTrain model families under shared splits.
  4. EvaluationReport comparable metrics and regimes.
Forecasting pipeline, benchmark harness, and evaluation framework engineeringPython / Time-series / Model benchmarking
Lab / lab / Quant & Trading / ongoing

Quant R&D Platform & Rust-backed Services

A constraint-aware research platform connected repeatable signal evaluation to performance-critical services.

Constraint Backtests must resist overfitting and measure robustness.

Select a signal node to inspect its operating principle.

  1. Research dataBuild repeatable feature inputs.
  2. BacktestingModel risk constraints explicitly.
  3. EvaluationMeasure robustness and resist overfitting.
  4. Rust servicesHarden performance-critical paths.
Research harness, backtesting primitives, and performance-oriented service engineeringRust / Python / Backtesting
Lab / open-source / Open Source / ongoing

LiteratureAtlas

An offline-first research atlas turns local PDFs into summaries, embeddings, maps, Q&A, and analytics.

Constraint PDF ingestion and generated outputs must remain local.

Select a signal node to inspect its operating principle.

  1. PDF ingestionKeep source material local.
  2. Local intelligenceSummarize and embed on device.
  3. OutputsPersist inspectable local artifacts.
  4. AtlasConnect maps, Q&A, and analytics.
SwiftUI application, local intelligence workflow, analytics, and optional accelerationSwiftUI / On-device LLM / Python
Lab / open-source / Open Source / ongoing

Codebase Combiner

A workspace becomes a filtered, prompt-ready artifact with predictable size and token estimates.

Constraint LLM context needs the right files in the right order at a predictable size.

Select a signal node to inspect its operating principle.

  1. WorkspaceDiscover candidate source files.
  2. SelectionApply glob and extension filters.
  3. SizingEstimate prompt tokens before export.
  4. ArtifactGenerate one prompt-ready document.
VS Code extension, macOS SwiftUI application, filtering, and prompt artifact toolingVS Code Extension / Node/JS / SwiftUI
Lab / open-source / Open Source / fork

Quants Lab (Research Base)

A forked research base provides notebooks, backtesting scaffolding, labeling, optimization, and visualization for experiments.

Constraint The repository is a forked base rather than an original RSI Tech project.

Select a signal node to inspect its operating principle.

  1. ResearchOrganize repeatable experiments.
  2. LabelingPrepare strategy evaluation inputs.
  3. BacktestingRun experiments against shared scaffolding.
  4. VisualizationInspect market and result data.
Use as a research sandbox for strategy experimentsPython / Notebooks / Backtesting