Skip to content

Mahesh Sai Kandula

I turn messy data intodecisions you can trust.

I work the full length of the pipeline: schema design, ETL, data-quality engineering, SQL business marts, modelling and BI, so that the numbers people act on hold up when someone checks them.

Data AnalyticsData EngineeringMachine Learning

Explore my workResume

LocationSydney, Australia

FocusAnalytics · Analytics Engineering

EducationMSc Data Science, Macquarie University

Open to opportunities

Tools and technologies I work with

  • Python
  • SQL
  • PostgreSQL
  • pandas
  • NumPy
  • ETL
  • Data modelling
  • Tableau
  • Power BI
  • matplotlib
  • scikit-learn
  • XGBoost
  • SHAP
  • TensorFlow / Keras
  • OpenCV
  • PyVista
  • Git
  • Jupyter
  • Jira
  • Excel
  • AWS (fundamentals)

How I work

From raw records to a decision.

Four stages, one continuous pipeline. Most of my projects run the whole length of it.

  1. 01Engineer

    Schema design, ETL and data-quality checks that flag bad records instead of quietly dropping them, so the audit trail survives.

    PostgreSQL · Python · SQL

  2. 02Analyze

    SQL business marts and exploration that turn clean tables into an answer a team can actually act on.

    SQL · pandas · Excel

  3. 03Model

    Supervised and deep learning to find the signal, then explainability to prove the model found the right one.

    scikit-learn · XGBoost · SHAP · TensorFlow

  4. 04Visualize

    Dashboards and visual systems that make a complicated result legible in a glance.

    Tableau · Power BI · matplotlib

Selected work

Four systems, built end to end.

04 case studies

  1. 01 · Analytics Engineering
    PayScope Tableau chargeback-analysis dashboard with KPI tiles, chargebacks by reason, dispute outcomes, and a chargeback amount trend line
    01 / 04Payments · Risk Analytics

    PayScope: Digital Payments Risk & Merchant Analytics Platform

    End-to-end payments analytics and risk monitoring on PostgreSQL, from raw data to live Tableau dashboards.

    • PostgreSQL
    • SQL
    • Python
    • Tableau
    100K+ transactions · 10K customers · 1.2K merchantsView case study
  2. 02 · Machine Learning
    SHAP beeswarm summary plot showing host_scale_20+, amenities_count, and number_of_reviews as the strongest drivers of predicted guest satisfaction
    02 / 04Hospitality · Product Analytics

    Sydney Airbnb Guest Satisfaction Classifier

    Supervised ML pipeline that finds what actually drives guest satisfaction while debunking a misleading signal.

    • Python
    • scikit-learn
    • XGBoost
    • SHAP
    17,730 Sydney listings · Sep 2025 scrapeView case study
  3. 03 · Data Visualisation
    3D lung mesh rendered by the pipeline for Case 01: advanced disease, progression score 0.809, deep fibrotic red with strong deformation
    03 / 04Healthcare · Medical Imaging

    Organ Disease Progression Visualizer: Imaging-Driven 3D Lung Model

    A deterministic pipeline that turns chest CT images into a live, signal-driven 3D lung visualisation.

    • Python
    • OpenCV
    • PyVista
    • Trimesh
    • NumPy
    8 clinical CT cases · scored 0.218–0.809View case study
  4. 04 · Computer Vision
    Normalised confusion matrix for the custom CNN, with per-class accuracy from 87% to 99% across eight tissue types
    04 / 04Histopathology · Deep Learning

    Colorectal Cancer Tissue Classification (Deep Learning)

    A custom CNN that beat VGG16 transfer learning at classifying eight histopathology tissue types.

    • TensorFlow/Keras
    • scikit-learn
    • NumPy
    5,000 histology images · 8 tissue typesView case study

What's next

Have an interesting data problem?Let's build something useful.

I'm looking for a junior data analyst or analytics engineer role in Sydney. If you have messy data and a decision waiting on it, I'd like to hear about it.

maheshsaikandula@gmail.com+61 451 138 996Resume

Open to opportunitiesSydney, Australia