4D Analytics Lab

A recruiter can open your repositories but still struggle to understand the engineering inside them. Static portfolio copy cannot answer follow-up questions or prove where a claim came from. Generic chatbots sound impressive until they start guessing. The challenge was to build a grounded assistant that explains real work using real code. Every answer needed evidence, traceability, and a delivery style that feels native to the portfolio.

Returned devices still contained physical SIM cards that needed to be identified and deactivated. Operators had to read tiny printed SIM identifiers and record them manually. That process was slow, vulnerable to transcription errors, and difficult to scale. I built a conveyor capture workflow, fine-tuned its OCR recognizer, and connected selective AI verification to auditable operator review.

Inventory operations ran across spreadsheets and ad-hoc apps. Data drift and duplicate records hurt counting, replenishment, and audits. Licensing changes pushed costs up and complicated scaling. Teams needed clarity: low-code speed vs custom control and TCO. Decision framework to model cost, risk, and vendor lock-in.

Most personal trading systems mix research, inference, execution, and risk into one process. That makes them easy to prototype and hard to trust under runtime stress. I wanted a platform where the part that thinks and the part that acts are deliberately separated. The challenge was to combine ML-driven directional trading, statistical arbitrage, and independent risk controls in one architecture. It needed to read like quantitative systems engineering, not trading hype.

Weekly uptime reporting was manual, repetitive, and too easy to get wrong. Teams had to capture availability across clients and locations, log outage comments, and still build a stakeholder PowerPoint by hand. That consumed hours of admin work every reporting cycle. The business needed one workflow for capture, insight, and presentation output. The challenge was to save time without compromising reporting accuracy.

Ever noticed this? Same kitchen. Same ingredients. One person somehow commits an atrocity against the human tongue while another turns those exact items into a meal that pulls the whole family to the table.