Attendees following a talk at IndabaX Eswatini 2026
IndabaX Eswatini logo
Deep Learning IndabaX · 8th edition

IndabaX
Eswatini

From community to intelligence.

The 8th Deep Learning IndabaX in Eswatini, a local gathering of the pan-African Deep Learning Indaba movement. This year, six speakers took the room from competing on Zindi, the foundations of machine learning and a hands-on practical to AI agents and the ecosystem Eswatini needs to put AI to work, alongside poster presentations of research from our community.

Talks
06
Speakers
06
Resources
06

Photos © Central Bank of Eswatini

Browse the archive

Year2026
LocationEswatini
1 edition so far
01

The Talks

From the basics of machine learning to agents that pause for a human, plus the ecosystem that makes all of it useful to Eswatini. Pick a speaker, or step through them in order.

Talk 01 of 06

Robert Selemani speaking at IndabaX
HackathonsTalk 01

Competing on Zindi: From Hackathons to an AI Career

Robert Selemani· Co-founder & Managing Director, AI Eswatini · AI Platforms Engineer, Cassava AI

Robert introduced Zindi: where and how it started, where it is today, and his own journey with the platform as a Zindi Ambassador. He then walked participants through the platform itself, from joining the hackathon and navigating its tabs to creating teams and making submissions.

Building on the Hackathon Preparation Webinar and the testimony of last year's winner, Robert set out how to approach the hackathon with a winning spirit: understand the problem statement, gain domain knowledge, do the research, iterate and test different approaches, and collaborate. He closed with his own story. Zindi competitions built the portfolio that landed him his first data role, and Zindi's job placements led to his current position as an AI Platforms Engineer at Cassava Technologies, under Cassava AI. Competing, and being part of a network of over 100,000 data enthusiasts, can transform a career.

Key takeaways

  1. 01Understand the problem statement before touching the data.
  2. 02Build domain knowledge and do the research.
  3. 03Iterate and test different approaches.
  4. 04Collaborate: teams are part of the platform for a reason.
  5. 05Competitions build a portfolio, and the network opens doors to real roles.

A full house, ready to build.

IndabaX Eswatini brought together students from a wide range of tertiary institutions and backgrounds, alongside researchers, practitioners and partners.

The full IndabaX Eswatini 2026 audience standing together in the hall, with guests and organisers seated in the front row
02

Posters

Alongside the talks, members of our community presented their research as posters: early work on problems that are specific to Eswatini.

NLP · Fraud detectionResearch in progress

Detecting “Facata”

An NLP-based and machine learning early warning system for mobile money scams in Eswatini

Facata is a common mobile money scam in Eswatini: someone claims to have sent money by mistake, then pressures the victim into sending it back or sharing a PIN. Nompendulo's research trains models to flag these messages in siSwati, English and the mix of both that people actually text in, before a victim responds.

reported cases, Apr 2025 to Mar 2026
605
lost by victims
E6.27M

The approach

  1. 01Compares a TF-IDF baseline, an explainable feature-based model and an African-language transformer.
  2. 02Builds a local, labelled dataset from consented messages, ESCCOM advisories and reviewed synthetic texts.
  3. 03Removes or hashes PINs, OTPs and numbers before any training.
03

Resources

Pick up where the sessions left off: work through the practicals in your browser, or view and download the slides and posters.

Repository

Machine learning basics practical

The notebook from Ayanda Khumalo's session: the ML workflow, preparing data, and training and evaluating a first model with scikit-learn.

View on GitHub
Notebook

Getting started with Colab & Kaggle

A companion notebook for running the practical in the browser with no local setup, and for continuing on Kaggle.

View on GitHub
Slides

Building the AI Ensemble

Samkelo Msibi's slides: people, language, robotics, quantum and applied AI.

View on GitHubDownload
Slides

From Community to Intelligence

Bandile Malaza's slides on building Eswatini's AI ecosystem for strategic decision-making.

View on GitHubDownload
Poster

Detecting “Facata”

Nompendulo Silombo's poster on an early warning system for mobile money scams in Eswatini.

View on GitHubDownload
Poster

Cash flow verification for MSMEs

Bongumenzi Makhanya and Nkosiyenzile Mantimakhulu's poster on business profiles and financial-stability scoring.

View on GitHubDownload
All 2026 presentations and posters on GitHub

The next IndabaX

Speak, build or learn with us.

Every talk on this page came from our community. Join to hear about the next edition first, or get involved through one of our programs.

Follow IndabaX Eswatini

Photos, highlights and news from the event, on the official IndabaX Eswatini pages.