Work & case studiesEvidence before claims.

Every project,
open to inspection.

Explore the questions, decisions, and artifacts behind our analysis, prototypes, concepts, and engineering methods.

Each entry identifies what was made, its source, and its limitations.

Explore the work Featured independent analysisCyclistic ridership

Independent analysis

From a ride record
to a clearer question.

Explore the Cyclistic source figures, model and disclosed assumptions.

Inspect the case study

Python repository artifact · source run unverified

Work, in motion.

Five perspectives on a technical problem. Follow the work, then inspect the evidence and its boundaries.

Browse the complete index

Applied AI evaluationPublished engineering method

Concept preview · evaluation method

Applied AI & Automation

Production RAG Validation Suite

A published evaluation method for testing retrieval quality, groundedness, and failure boundaries before a knowledge assistant is released.

Hybrid retrievalCross-encoder rerankingGroundedness evaluation

Workflow automationInterface concept

Concept interface · illustrative flow

Systems & Software

Automated Incident Dispatcher

A concept interface for webhook intake, signature checks, and routing an incident to an on-call owner.

Interface designWebhook flowRouting contracts

Data engineering & BIIndependent portfolio project

Python repository artifact

Source figurePython analysis artifact

Python chart of daily bike ride counts, with date on the horizontal axis and ride count on the vertical axis.

Daily ride counts · Python analysis artifact. The chart’s source run is separate from the BI demonstration export.

Data Analytics & BI

Cyclistic Ridership Analysis

Python analysis artifacts, a dimensional BI model, and a financial scenario explorer.

Python / PandasPower BI / DAXStar Schema

Systems & cloudInteractive system prototype

Illustrative system map

Systems & Software

Cloud Architecture Visualizer

An interactive system map for following data through a knowledge assistant, API synchronization, and a BI pipeline.

TypeScriptReactNext.js

Operational engineeringPublished decision framework

Illustrative decision framework

Systems & Software

Workflow Automation Feasibility Model

A decision framework for choosing between deterministic scripts, queued webhooks, AI assistance, and human review.

Process mappingRisk assessmentMaintenance analysis

Projects, prototypes
& published methods.

6 entries

Choose a project

Independent portfolio projectData engineering & BI

Cyclistic Ridership Analysis

Python analysis artifacts, a dimensional BI model, and a financial scenario explorer.

Python / Pandas
Power BI / DAX
Star Schema
Scenario modeling

Read the case study Python source BI source

Question, contribution & scope

The question: How do casual and member usage patterns differ, and how could an analytical model support a membership-conversion decision?
The approach: Separate data preparation, exploratory analysis, dimensional modeling, and scenario assumptions so each can be inspected on its own.

A modular Python pipeline, analysis figures, a one-fact/six-dimension BI structure, DAX definitions, and a browser-based scenario model.

The BI export contains 61,973 demonstration records. The repositories expose the analytical workflow and modeling logic, without establishing realized commercial results.

Scope: Independent work, not a client engagement. Data-generation code is present in the source; historical provenance and the original run behind the Python figures are not independently established. Financial outputs are hypothetical.

.work-browser__search,.work-browser__mobile,.work-browser__toolbar,.work-browser__matches{display:none!important}.work-browser__directory{display:block!important}.work-browser__list{max-height:360px}.work-browser__panels:has(:target)>.work-browser__panel:not(:has(:target)){display:none!important}.work-browser__panel:has(:target){display:block!important}

Trace the system.
Inspect the boundary.

The architecture explorer is a working frontend prototype. Follow a stage to see its inputs, outputs, and failure boundaries.

.arch-system__toolbar,.arch-canvas,.arch-deck{display:none!important}

Architecture workflow reference

Explanatory prototype. These workflows describe contracts and failure paths; they do not run production infrastructure.

Grounded Knowledge Assistant

Ground language models in authoritative corporate documents, enforce citation boundaries, and maintain a deterministic human escalation path for low-confidence inference.

Document Intake

Parse heterogeneous enterprise documents, extract high-fidelity text and tabular structures, and strip non-authoritative boilerplate before indexing.

Contract: Cryptographic SHA-256 hash verified; metadata normalized before parsing.
Failure mitigation: Corrupt or encrypted payloads quarantined to Dead-Letter Queue.
Output artifact: Sanitized markdown AST with tabular preservation

Semantic Chunking

Structure unstructured prose into hierarchical parent-child passage chunks, enabling exact keyword retrieval alongside dense semantic vector similarity.

Contract: Strict 512-token semantic boundaries with 10% sliding overlap.
Failure mitigation: Oversized tables decomposed into relational dimension facts.
Output artifact: Dual index: Dense embeddings + BM25 lexical inverted index

Guardrailed Retrieval

Query-time hybrid retrieval merges vector proximity with lexical matching, followed by a cross-encoder model that scores passage relevancy before prompt injection.

Contract: Reciprocal Rank Fusion (RRF) merges top-40 dense and sparse candidates.
Failure mitigation: Low lexical overlap triggers query expansion fallback.
Output artifact: Top-5 reranked context passages with similarity scores

Grounded Handover

Present a draft with source references, explicit limitations and a review path for unsupported or ambiguous requests.

Contract: Every declarative sentence carries an interactive source page anchor.
Failure mitigation: Out-of-domain queries trigger graceful fallback disclosure.
Output artifact: Streaming SSE response with verifiable citations & audit log

Resilient Event Broker

Connect mission-critical external APIs and internal microservices under real network conditions, handling network partitions, duplicate webhooks, and schema drift safely.

Inbound Webhook

Receive partner webhook events at edge endpoints with cryptographic authenticity checks, validating timestamp freshness to prevent replay attacks.

Contract: HMAC SHA-256 cryptographic signature verified before body parsing.
Failure mitigation: Invalid signatures are rejected before the event enters the processing queue.
Output artifact: Authenticated raw event payload with ingress timestamp

Asynchronous Queue

Decouple reception from processing, with idempotency checks and reconciliation to reduce the risk of duplicate downstream effects.

Contract: UUIDv5 idempotency key cached in distributed key-value store for 72h.
Failure mitigation: Duplicate event IDs acknowledged immediately with HTTP 200 without reprocessing.
Output artifact: Deduplicated FIFO queue message buffer

Transform & Validate

Transform external partner schemas into your internal canonical representation, isolating vendor-specific formatting changes from your core business logic.

Contract: Zod runtime schema validation against canonical enterprise entity model.
Failure mitigation: Schema drift anomalies routed to schema-mismatch quarantine topic.
Output artifact: Normalized domain event with strongly typed relations

Downstream Dispatch

Write transformed entities to target databases and third-party APIs with atomic transaction boundaries and automated daily reconciliation auditing.

Contract: Bounded exponential backoff with decorrelated jitter (max 5 retries).
Failure mitigation: Unrecoverable downstream timeouts trigger circuit breaker.
Output artifact: Transactional write confirmation & reconciliation receipt

BI & Scenario Engine

Transform disparate operational logs into an unambiguous Star Schema data model with auditable DAX measures and dynamic what-if simulation parameters.

Extract & Profile

Profile operational tables before transformation, detecting upstream database migrations, missing foreign keys, or unexpected datatype alterations early.

Contract: Null checks, distribution skew, and referential integrity profiled on ingest.
Failure mitigation: Malformed records flagged and written to profiling exception log.
Output artifact: Staged raw tables with verified row counts and column profiles

Star Schema Modeling

Build normalized dimension tables and narrow fact tables with integer surrogate keys, ensuring lightning-fast columnar compression and aggregation.

Contract: Strict 1-to-many single-direction relationships between dimensions and facts.
Failure mitigation: Many-to-many bridge tables prohibited to guarantee predictable filter propagation.
Output artifact: Optimized Parquet analytical warehouse tables

Measure Definition

Implement explicit DAX measures with strict filter context boundaries, eliminating conflicting metric definitions across departments.

Contract: All business KPIs defined once as explicit measures; zero implicit columns.
Failure mitigation: Calculations validated against raw ledger aggregates with automated unit tests.
Output artifact: Version-controlled tabular semantic model with calculation logic

Decision Surface

Deliver an executive decision surface where operators can toggle pricing, conversion, and capacity assumptions in real time to forecast operational impact.

Contract: Disconnected parameter tables simulate dynamic revenue & volume sensitivities.
Failure mitigation: Boundary inputs clamped to mathematically valid business domains.
Output artifact: Interactive decision dashboard with dynamic scenario simulator

Ground language models in authoritative corporate documents, enforce citation boundaries, and maintain a deterministic human escalation path for low-confidence inference.

Illustrative design / no live system connection

Human Review Escalation

Passages with cross-encoder rerank confidence < 0.78 route to human triage with exact retrieved document context.

Contract invariant

Cryptographic SHA-256 hash verified; metadata normalized before parsing.

Failure mitigation

Corrupt or encrypted payloads quarantined to Dead-Letter Queue.

Output artifact

Sanitized markdown AST with tabular preservation

Stage 1 of 4

Have a problem that does not fit a category?

Most real systems cross more than one discipline. Bring the constraint and we will map the smallest reliable path through it.

Discuss a project See the approach

Every project,
open to inspection.

From a ride record
to a clearer question.

Work, in motion.

A plan for evaluating
the answer.

Production RAG Validation Suite

Automated Incident Dispatcher

Cyclistic Ridership Analysis

Cloud Architecture Visualizer

Choose the right
automation boundary.

Workflow Automation Feasibility Model

Projects, prototypes
& published methods.

Cyclistic Ridership Analysis

Cloud Architecture Visualizer

Automated Incident Dispatcher

Production RAG Validation Suite

A plan for evaluating
the answer.

Workflow Automation Feasibility Model

Choose the right
automation boundary.

Executive BI & Star Schema Blueprint

One model.
Explicit definitions.

Trace the system.
Inspect the boundary.

Follow the architecture.

Document Intake

Contract invariant

Failure mitigation

Output artifact

Have a problem that does not fit a category?

Bring the right
pieces together.

Work, in motion.

Projects, prototypes& published methods.

Trace the system.Inspect the boundary.

Follow the architecture.

Document Intake

Contract invariant

Failure mitigation

Output artifact

Have a problem that does not fit a category?

Bring the rightpieces together.

Projects, prototypes
& published methods.

Trace the system.
Inspect the boundary.

Bring the right
pieces together.