Projects

Systems, Not Screenshots.

Engineering systems built around real problems, explicit architecture and verifiable results.

Each project documents its operational boundaries and relevance to similar product or infrastructure work.

01 / Flagship systems

Primary Portfolio Collection

Versioned release

Product intelligence platform

RetentionOS

Analytics · Experimentation · Churn ML · Platform Engineering

Problem fit: SaaS analytics, retention intelligence, churn analysis, experimentation and internal product-data tooling.

System

A reproducible product-intelligence platform joining behavioral events, SQL analytics, retention and funnel analysis, calibrated churn prediction and governed A/B testing.

Strongest proofLive end-to-end workflow, v1.0.0 release and 167 backend tests passed

10KSynthetic users
~238KProduct events
167 passedBackend tests
v1.0.0Release
PythonFastAPIPostgreSQLRedis
Versioned release

Reliable infrastructure for AI/ML services

Resilient Distributed Runtime

Distributed Systems · Reliability · Kubernetes · AWS

Problem fit: backend reliability, observability, container/cloud delivery, distributed coordination and resilient AI/ML infrastructure.

System

Correctness-first infrastructure combining bounded execution, cluster membership, causal CRDT replication, durable recovery, mTLS, observability, chaos testing and reproducible delivery.

Strongest proofThree-node Kubernetes verification with failure recovery and 439 tests passed

439 passedAutomated tests
3 nodesLocal cluster
v0.7.0Release
EKS verifiedCloud lifecycle
PythonCRDTsetcdmTLS
Versioned release

Synthetic ML governance case study

CreditScoreV4 ML Governance

Model Risk · Responsible AI · MLOps · Release Safety

Problem fit: ML serving, monitoring, data quality, drift detection, governed releases and production-readiness reviews.

System

A deterministic governance system testing whether unsafe data and model changes are blocked through independent quality, drift, fairness, integrity, promotion, rollback and incident controls.

Strongest proofTagged v1.0.1 release with independent fail-closed controls

v1.0.1Release
Fail-closedGovernance
Canary + rollbackRollout
AWS gatedDeployment
PythonFastAPIGreat ExpectationsSHAP
Versioned release

AI-assisted portfolio research system

AXIOM Portfolio Intelligence

Constrained MPT · FinBERT · FAISS · Evidence-grounded RAG

Problem fit: quantitative research platforms combining analytical computation, document retrieval and source-grounded AI explanations.

System

A research system that evaluates constrained portfolio strategies and produces source-grounded explanations without allowing generated text to change allocations.

Strongest proofEqual weighting outperformed the optimized strategy on return and Sharpe in the documented walk-forward evaluation

84Automated tests
v1.0.0Release
Walk-forwardEvaluation
Equal weightStronger baseline
PythonFinBERTFAISSLangChain

Temporary research endpoint; not a production financial service.

02 / Earlier work

Applied Machine-Learning Projects

Smaller learning projects showing model selection, evaluation and problem-specific safeguards.

Applied ML

Applied ML learning project

Audi Car Price Prediction

Regression · EDA · Model comparison

System

An applied regression workflow covering exploratory analysis, preprocessing, model comparison and R²/MAE evaluation on Audi vehicle data.

RegressionTask
R² + MAEEvaluation
PythonScikit-learnRandom ForestCatBoost
Focused learning project
Applied ML

Applied ML learning project

Credit Card Fraud Detection

Imbalanced classification · Evaluation · Explainability

System

An applied classification project focused on leakage-safe preprocessing, imbalanced evaluation, threshold behavior and clearer model explanations.

ClassificationTask
ImbalanceFocus
PythonScikit-learnXGBoostSMOTE
Focused learning project
Repository not published
Applied ML

Applied ML learning project

Parkinson's Disease Prediction

Voice features · Classification · Model comparison

System

An educational classification workflow using voice measurements and comparing Logistic Regression, SVM, Random Forest and KNN.

ClassificationTask
Precision + recallEvaluation
PythonScikit-learnSVMRandom Forest
Focused learning project