Ng'andu Fad Sikazwe

Experience behind the work

FigureFlow

Consulting technical lead for a Swedish financial SaaS company since 2025. I lead engineering across accounting imports, reconciliation, reporting and AI workflows, with access controls, human review and audit records. My financial software experience

11 years

Building and supporting software since 2015, from the first client meeting to support after launch.

99.5%

Precision on accepted decisions in an AI routing evaluation of more than 10,000 test cases. The system could decline and pass uncertain cases to a person. This measures the decisions it accepted, not every request. The rule is short enough to read:

def route(decision, threshold=0.9):
    """Act when the model is confident. Otherwise, ask a person."""
    if decision.label == "none" or decision.confidence < threshold:
        return send_to_review(decision)
    return apply(decision)

14:25

The time in Lusaka, Zambia now, UTC+2. I work with teams in Europe and Africa.

Stack

Python, Django, PostgreSQL, Redis · Celery, Vue 3, TypeScript, React · React Native, Node.js. I choose tools to fit each problem.