Grouped by discipline · each linked to a project

Skills.

01
AI & ML Systems
Models taken from data to a measured, deployed service: classification, forecasting, computer vision and physics-informed networks.
02
AI Solutions
LLM and RAG products that answer from a business's own content, cite sources and hand off to people when unsure.
03
AI Agents
Assistants and automations that retrieve, decide and act: grounded answers, data-aware portal assistants, scheduled check-ins and escalations.
04
Software Development
Full-stack web, iOS and Android products with role-based access, payments, tests and production deployment.

Tools and languages, by discipline

How I build

01
Frame
Define the user problem and success metric before implementation.
Safira: grounded answers or human hand-off
02
Data
Collect, clean, chunk and version the source data.
Safarnama content indexed in pgvector
03
Model
Start with the simplest model that meets the target, then tune.
Road Health: random forest on five features
04
Evaluate
Measure against labelled test sets and automated tests.
Safira API: 11 tests · Road Health: 5
05
Deploy
Ship resilient APIs with queues, retries and monitoring.
Transactional outbox with SKIP LOCKED
06
Monitor
Log real usage and improve on observed failures.
Level Up chatbot feedback loop
07
Responsible AI
Grounding, citations, audit trails and accessibility by default.
Citations · human hand-off · audit log

Contact