How is this course delivered?

This is a tutor-led course delivered live online, taught in English (Uganda). Your tutor guides you through the curriculum live rather than leaving you to self-study pre-recorded videos.

Will I get a certificate?

Yes. You'll receive a course 26: speech recognition & voice ai for local languages certificate after completing the course requirements.

What do I need to know before I start?

You should have: Having Successfully learnt Course:NLP for African Languages.

How long does the course take, and what does it cost?

The course runs for 8 weeks and costs UGX 700,000. Paying for one programming course also unlocks 25% off your next one.

What do I need to have ready to join?

You'll need: Computer with microphone; Headphones; Stable internet connection; Python 3; Jupyter Notebook or Google Colab; Code editor such as Visual Studio Code; Audio editor such as Audacity; Access to a speech-to-text and text-to-speech API or open-source speech model.

Can I get help if I get stuck?

Yes. Enrolled students can book one-on-one tutor sessions any time through Get Help, in addition to regular class time.

Programming University · General PublicSpeech Recognition & Voice AI for Local Languages

Machine Learning

University · General Public

Speech Recognition & Voice AI for Local Languages

Duration

8 weeks

Investment

UGX 700,000

Certificate

Included

Teaching

Live online

Enroll in this tutor-led course Get Help

Program Introduction

This advanced speech recognition and Voice AI course is designed for developers, data scientists, researchers and digital-product teams in Uganda, East Africa and across Africa. Learners work with audio data, speech-to-text, text-to-speech, responsible local-language datasets, accent adaptation and inclusive IVR/USSD design, then build a small farming-advice prototype that responds to spoken Luganda questions.

Key Features & Benefits

• Uganda and East Africa context • Hands-on speech-to-text and text-to-speech • Responsible local-language dataset preparation • Local-accent and low-resource model adaptation • Inclusive IVR and USSD voice-interface design • Luganda farming-advice capstone

Real-World Applications

• Voice-enabled agricultural advisory and farmer helplines • Local-language customer support and call-centre transcription • Accessible IVR and USSD services for low-literacy users • Radio and media transcription or monitoring • Voice interfaces for education, public services and community information • African-language voice assistants and smart-device controls

Course outline and learning expectations

This is a tutor-led course. The outline shows what your tutor will cover; teaching materials and examinations are provided directly to enrolled students.

Speech RecognitionVoice AIAutomatic Speech RecognitionASRSpeech-to-TextText-to-SpeechTTSLugandaUgandan LanguagesAfrican LanguagesEast AfricaLocal Language AILow-Resource LanguagesIVRUSSDAudio DatasetsMachine LearningNatural Language ProcessingNLPUganda AI Course

Teaching format

Live online

Teaching language

English (Uganda)

Intended learners

University · General Public

What you will learn

Explain how speech recognition systems process audio data
Use speech-to-text tools and APIs to transcribe recorded or live speech
Build and evaluate a basic text-to-speech workflow
Assess speech-model performance across local languages, accents and recording conditions
Plan and prepare a responsibly collected local-language audio dataset
Adapt or fine-tune a speech model for a selected local language or accent
Design an inclusive voice interface that combines IVR, USSD and speech
Build and demonstrate a small voice-enabled tool that responds to spoken Luganda questions

Modules

1
Speech Recognition Foundations
Understand audio as data and the main stages of automatic speech recognition
WaveformsSampling rate and bit depthSpectrograms and acoustic featuresASR pipelineAudio quality and noise
2
Speech-to-Text Tools and APIs
Use cloud and open-source speech-to-text tools and compare their output on African accents
API authentication and requestsAudio formatsBatch and streaming transcriptionConfidence scoresWord error rateTesting Ugandan and East African accents
3
Text-to-Speech Basics
Build and evaluate simple text-to-speech workflows for local-language applications
Text normalizationPronunciation and phonemesVoice synthesis pipelineSSML basicsNaturalness and intelligibility testing
4
Local Languages and Accent Challenges
Analyse why existing speech models may underperform on low-resource African languages and diverse accents
Data scarcityCode-switchingDialect and accent variationDomain mismatchNoise and recording conditionsFairness and bias
5
Responsible Local-Language Audio Datasets
Plan, collect, label and document speech data with consent and community safeguards
Use-case definitionSpeaker consentSampling and representationRecording protocolsTranscription and annotationQuality controlPrivacy and dataset documentation
6
Adapting Speech Models for Local Accents
Adapt or fine-tune a speech model and measure results on a selected Ugandan or African language
Transfer learningData preparationTrain-validation-test splitsFeature extractionFine-tuning workflowError analysisModel evaluation
7
Voice Interfaces for Low-Literacy Users
Design inclusive voice services that combine IVR, USSD and speech for practical African contexts
Conversational flow designPrompt writingFallbacks and confirmationsIVR integrationUSSD hand-offAccessibilityOffline and low-bandwidth considerations
8
Capstone Voice-Enabled Tool
Build a small voice-based farming-advice IVR prototype that responds to spoken Luganda questions
Problem definitionLuganda question setSpeech-to-text integrationIntent or response logicText-to-speech outputUser testingDocumentation and demonstration