Level 1 · The investment process

Across the full investment workflow.

Eight stages turn raw data into portfolio weights and back into evidence. Each one has its own transformation, its own failure modes and its own systems. Select a stage to see what happens there and the work behind it.

QUANT INVESTMENT WORKFLOWDATA → RESEARCH → PORTFOLIO → EVIDENCE
The premise Every portfolio weight is the output of a data pipeline: identifiers resolved, prices adjusted, factors scored, forecasts optimized. I build the systems that make each step correct, reproducible and reusable.

Ideas create alpha. Infrastructure decides whether that alpha survives contact with production. Select a stage to see the transformation it performs and the systems I built for it.

Selected stage

What happens here

My experience

Technical layer

Feeds next

Level 3 · The data stack

From vendor feed to portfolio decision.

Four kinds of data, seven layers, one path to the portfolio. I have built, owned or rebuilt work at every layer.

Data
Decision
Analyst estimates
▶
Corporate actions
▶
Macro data
▶
Client restrictions
▶
AUTO

My part

LanguagesSQL (primary) · Python · R
Machine learningKNN · Random Forest · SVM · Naive Bayes · PCA · Regression · TF-IDF · Reranking transformers
PracticeAgile · CI/CD · Model development · Predictive analytics
NextWorkCase studies: what was built and led