Skip to content
View sairakbar's full-sized avatar

Block or report sairakbar

Block user

Prevent this user from interacting with your repositories and sending you notifications. Learn more about blocking users.

You must be logged in to block users.

Maximum 250 characters. Please don’t include any personal information such as legal names or email addresses. Markdown is supported. This note will only be visible to you.
Report abuse

Contact GitHub support about this user’s behavior. Learn more about reporting abuse.

Report abuse
sairakbar/README.md

hello! 🪼

This is where I keep the projects I build outside of work.

Most started with a question I couldn't stop thinking about. I usually build something before I convince myself I know the answer.

I'm slowly moving everything here, and adding new ideas as I go.

questions currently scratching my brain

  • When should a product trust its own model?
  • How do you know a good decision when the outcome is uncertain?
  • What makes an explanation genuinely useful?
  • Can we build systems that are both accurate and understandable?

current lists with some interesting stuff

✈️ the airport has feelings

Machine learning for the moments when airport plans meet operational reality.

[The Airport Gets Tired] (https://github.com/sairakbar/airport-gets-tired) . [Heathrow Connection Control Tower] (https://github.com/sairakbar/heathrow-connection-control-tower) . [Heathrow's Breaking Point] (https://github.com/sairakbar/heathrow-breaking-point) . [The Ghost Schedule] (https://github.com/sairakbar/ghost-schedule) .

💷 money, models & consequences

Responsible decision systems for credit, risk and financial services.

[Thin File] (https://github.com/sairakbar/thin-file) . [The Humility Engine] (https://github.com/sairakbar/humility-engine) . [Model Risk & Fairness Monitoring Framework] (https://github.com/sairakbar/model-risk-monitoring-framework) . [Explainable Loan Underwriter AI] (https://github.com/sairakbar/explainable-loan-underwriter-ai) .

🏘️ buildings that talk back

Predictive intelligence for safer homes, smarter investment and healthier assets.

[The Homes We Know Least About] (https://github.com/sairakbar/homes-least-known) . [Intelligent Construction Project Risk Prediction Platform] (https://github.com/sairakbar/Intelligent-Construction-Project-Risk-Prediction-Platform) . [Predictive Project Risk Intelligence Platform] (https://github.com/sairakbar/Predictive-Project-Risk-Intelligence-Platform) .

🧠 ai with self-doubt

AI systems that know when their memory, confidence or judgement may be wrong.

[The Memory Drift] (https://github.com/sairakbar/memory-drift) . [The Humility Engine] (https://github.com/sairakbar/humility-engine) . [Model Risk & Fairness Monitoring Framework] (https://github.com/sairakbar/model-risk-monitoring-framework) .

🏛️ fix the system, not the people

Organisational intelligence that finds friction without blaming the workforce.

[The Council's Second Job] (https://github.com/sairakbar/councils-second-job) .

🌱 carbon, but causal

Testing whether sustainability interventions genuinely change behaviour.

[The Carbon Butterfly Engine] (https://github.com/sairakbar/carbon-butterfly-engine) .

Pinned Loading

  1. heathrow-breaking-point heathrow-breaking-point Public

    Simulates a busy airport hour-by-hour to find exactly which part breaks first under pressure, then suggests the smallest fix. Built with forecasting, queueing theory, and optimisation.

    HTML 1

  2. memory-drift memory-drift Public

    A memory system for AI that fades over time instead of either remembering everything forever or panicking at the first inconsistent thing someone says.

    Python 1

  3. carbon-butterfly-engine carbon-butterfly-engine Public

    Does switching to rail actually cut down how much someone flies, or is that just a nice story correlation tells us? Built to find out properly, with propensity score matching, a validation check ag…

    Python 1

  4. ghost-schedule ghost-schedule Public

    Predicts what happens to passengers after their flight is cancelled, where they go, when they'll actually resolve, and what it costs the airport. Classification vs. a transparent benchmark, surviva…

    HTML 1

  5. homes-least-known homes-least-known Public

    Most housing risk models ask "which homes will fail?" This one asks "which homes are we dangerously confident about?" Separates genuine low-risk from low-risk-because-nobody-checked, using ensemble…

    Python 1

  6. explainable-loan-underwriter-ai explainable-loan-underwriter-ai Public

    Human-in-the-loop credit decisioning platform with SHAP explainability, Fairlearn bias auditing, MLflow tracking, and a counterfactual recourse engine that tells declined applicants exactly what to…

    Python 1