resume
|
SEYDOU DIALLOSoftware Developer | AI/NLP Enthusiast | Linux Administrator | Community Builder
|
PROFILE
Software developer and AI/NLP researcher building speech and language technology for African languages - ASR, TTS, machine translation, and LLMs. Founder of MALIBA-AI and co-founder of DJELIA, where I take models from data collection through production deployment. First-author work on Bambara language technology published at peer-reviewed venues, with open-source models released on Hugging Face and developer SDKs in production.
EDUCATION
| Dakar American University of Science and Technology | Dakar, Senegal |
| Computer Science | 2024 - Present |
| Ecole Normal D’Enseignment Technique et Professional | Bamako, Mali |
| Electronic Engineering | 2023 - 2024 |
| Technical Baccalaureate in Electronic Engineering | Bamako, Mali |
| Industrial Sciences and Technologies | July - 2022 |
First National in Malian Technical Baccalaureate session
Specialization: Electronic Engineering
PUBLICATIONS
BambaraMLLM: A Unified Multilingual Multimodal Large Language Model for Comprehensive Bambara Language Processing
Seydou Diallo, Allahsera Auguste Tapo, Kevin Assogba, Christopher Homan - First author
- Unified multimodal model covering ASR, machine translation, text generation, and TTS for Bambara in a single framework
- Built a linguistically informed instruction-data framework injecting Bambara morphology, syntax, and cultural context to generate 2M+ conversational instruction examples
- Paper
Where Are We At with Automatic Speech Recognition for the Bambara Language?
Seydou Diallo, Yacouba Diarra, Mamadou K. Keita, Panga Azazia Kamaté, Adam Bouno Kampo, Aboubacar Ouattara - First author
- Introduced the first standardized ASR benchmark for Bambara, built on professionally recorded Malian constitutional text with manual segmentation, alignment, and quality review
- Evaluated 37 systems and released a public leaderboard for transparent Bambara speech-technology evaluation
- arXiv: 2602.09785
InstructLR: A Scalable Approach to Create Instruction Dataset for Under-Resourced Languages
Mamadou K. Keita, Sebastien Diarra, Christopher Homan, Seydou Diallo - Co-author
- Framework combining LLM generation, RAG-based correction, and human-in-the-loop validation to build high-quality instruction datasets for low-resource languages
- Produced the 50k-scale BambaraInstruct, ZarmaInstruct, and FulfuldeInstruct benchmarks
- arXiv: 2512.02213
GAIFE: Generative AI for Education (NAACL 2025 Findings) - Co-author
- Human-in-the-loop workflow using LLMs, machine translation, and linguistic tools to produce culturally grounded children’s reading materials in Bambara, a low-resource language
- Built a learning library of 174 illustrated books (~4,000 pages, 850+ original images), multiplying Bambara children’s content online roughly tenfold
- Pilot reading programs achieved a 67% reduction in children unable to read Bambara
-
Paper Library
PROFESSIONAL EXPERIENCE
Co-Founder | DJELIA
2024 – Present
- Own the full AI model lifecycle for Bambara - data collection, training, evaluation, and deployment
- Built and maintain production-grade ASR, TTS, MT, and LLM models for Bambara, served via cloud APIs
- Shipped Python and JavaScript SDKs so third-party applications can integrate Djelia’s language models
- Turned research-stage language models into reliable, documented services for local developers
Founder | MALIBA-AI Community
2025 – Present
- Founded and lead a community-driven AI initiative for indigenous Malian languages, owning research from data collection to final model
- Built and maintain ASR, TTS, LLM, and MT models across multiple Malian languages
- Released inference SDKs and serving infrastructure so developers and communities can integrate MALIBA-AI models directly
- Designed speech-first interfaces for primarily oral languages, expanding AI access in education, healthcare, agriculture, and local business
Software Developer & AI Engineer | Caytu Robotics
2024 – Present
- Built and maintain a scalable IoT ecosystem for remote control, real-time video streaming, and data monitoring
- Developed Caytu Avatar, a multimodal AI agent enabling natural-language interaction with any REST API, including the Caytu IoT ecosystem
- Created tools and SDKs for integrating AI services, automation workflows, and assistant features
Robotics Team Member | Malian National Team
2021, 2023
- Represented Mali in international robotics competitions
- Collaborated on complex engineering projects and problem-solving challenges
RESEARCH PROJECTS
Large Language Models
| MALIBA-LLM | Lead Developer | MALIBA-AI |
- Built MALIBA-LLM, the first open-source large language model for Bambara (Bamanankan), fine-tuned from google/gemma-3n-E2B-it
- Supports Bambara ↔ French/English code-switching, instruction following, translation, and Mali-specific knowledge
- Reduced validation loss to 0.4952 (93.4% lower than the 7.4595 starting point)
- Released on Hugging Face for education, translation, and conversational use: maliba-llm
Speech Synthesis
| Bambara Text-to-Speech | Lead Developer | MALIBA-AI |
- Open-source Bambara TTS system rated 4.2/5.0 for quality and 4.1/5.0 for naturalness by 10 native Bambara speakers
- Targets the digital exclusion affecting millions of non-literate adults in sub-Saharan Africa
- Released on Hugging Face - 8,712 downloads as of October 2025: MALIBA-AI/bambara-tts
MALIBA-TTS: Multilingual Speech Synthesis for Malian Languages | Lead Contributor
- Developed text-to-speech (TTS) models for six Malian languages: Bambara, Boomu, Dogon, Pular, Songhoy, and Tamasheq
- Based on Meta’s MMS-TTS and VITS, optimized for CPU inference and real-time synthesis
- Enables voice interfaces, education, and accessibility tools across Mali
- Promotes language inclusion and digital accessibility under the MALIBA-AI initiative
- Models: huggingface.co/MALIBA-AI/malian-tts
GAIFE: Generative AI for Education
Co-Author | Bill & Melinda Gates Foundation Grand Challenges | RobotsMali
- Co-developed a novel methodology combining generative AI and automatic translation to produce educational content at scale for Bambara, a low-resource language
- Co-author on the resulting peer-reviewed paper (NAACL 2025 Findings)
- Applied the methodology to build a library of 174 published illustrated books, contributing Bambara linguistic review for accuracy, fluency, and cultural relevance
-
Paper Library
Bayelemabaga: Bambara Machine Translation
Data Processing | RobotsMali | RIT Partnership
- Contributed to the first machine translation model for the Bambara–French language pair
- Handled extraction, cleaning, and alignment of the parallel corpora behind the model
- Project Details
- Dataset
Automatic Speech Recognition Systems
Wolof ASR Models | Lead Developer | CAYTU ROBOTICS
- Built whosper-large-v2 and whosper-large, open-source Wolof ASR models
- Handle multilingual input across Wolof, French, and English
- whosper-large-v2 reached 24% WER and 11% CER on a 9-hour evaluation set
- 3,893 downloads (whosper-large-v2) and 679 downloads (whosper-large) in first week
-
Models: whosper-large-v2 whosper-large
Bambara ASR Models | Lead Developer | MALIBA-AI
- Built the maliba-asr series for Bambara ASR
- maliba-asr-v2 reached 24.33% WER and 12.45% CER on a 3-hour evaluation set
- Handles Bambara–English code-switching
- Latest Model : bambara-asr-v1
Songhay ASR Model | | Lead Developer | MALIBA-AI
- Built songhay-asr-v1, an ASR model for the Songhay language
- Reached 16.58% WER and 4.63% CER on the test set
- Supports bilingual Songhay–French transcription
- Model: songhoy-asr-v1
IoT Assistant in Local Languages
Lead Developer | Caytu Robotics
- Built an IoT assistant operating fully in Wolof
- Designed a scalable IoT ecosystem with remote control and real-time streaming
- Added workflow automation and multilingual support
- Platform: caytu.link
- Open-sourced architecture at Ndabax Senegal 2024
OPEN SOURCE CONTRIBUTIONS
Software Development Kits (SDKs)
NabooPay Payment Gateway SDKs | Lead Developer & Maintainer
- Python SDK for Senegalese mobile money payments (Wave, Orange Money, Free Money)
- Node.js SDK in TypeScript with full type definitions and automatic retry logic
- Transaction management and cashout operations, with both sync and async support
-
Code Source: Python SDK Node.js SDK
Djelia AI Language Platform SDKs | Lead Developer & Maintainer
- Python SDK: Complete SDK for Bambara language AI (translation, ASR, TTS) with streaming support
- JavaScript SDK: Modern implementation for web and Node.js environments
- Features: Multi-language translation, real-time transcription streaming, natural TTS with speaker descriptions
-
Code Source: Python SDK JavaScript SDK
MALIBA-AI Bambara TTS SDK | Lead Developer & Maintainer
- Inference SDK for Bambara text-to-speech synthesis
- Production-ready inference system with 10 authentic speaker voices
- Optimized for educational platforms, accessibility tools, and voice interfaces
- Comprehensive documentation with usage examples and best practices
- Code Source: GitHub Repository
Infrastructure & Deployment Tools
Realtime Video/Audio WebRTC Streaming
- Enterprise-grade streaming solution containerized for easy deployment
- Enables low-latency cloud transmission from various sources
- Applications in remote monitoring, surveillance, telehealth, industrial IoT
-
Code source: GitHub Docker Hub
Wolof TTS Inference Server
- Deployment server for a Wolof speech synthesis system
- Makes voice technology available for Wolof, a primary language of Senegal
- Code Source: GitHub Repository
Adia_TTS Segmentation Algorithm
- Turned a research model into a production-ready system
- Removed character-length limits while preserving voice quality
- Merged into the official model documentation
-
Code Source: GitHub HuggingFace Discussion
Additional projects: GitHub Profile
PREVIOUS VENTURES
KAALISI | Founder
AI-powered investment evaluation platform
- Developed scoring models assessing startups on financial performance, environmental impact, pandemic resilience, and youth employment
- Aimed to democratize access to capital through intelligent matchmaking
JAYAS Drone | Co-Founder
Drone-as-a-Service platform
- Built ready-to-fly drones and service infrastructure
TECHNICAL SKILLS
Programming Languages: Python, JavaScript, C/C++, Bash, Java
AI/ML Frameworks: PyTorch, Hugging Face Transformers, Whisper, FastAPI, Unsloth, Axolotl
Technologies: Docker, WebRTC, IoT, AWS, GCP, Linux Administration, git, Github
NLP Specializations: Speech Recognition (ASR), Text-to-Speech (TTS), Machine Translation (MT), Large Language Model (LLM)
Languages: Bambara (Native), French (Fluent), English (Fluent)
COMMUNITY LEADERSHIP
MALIBA-AI - Founder & Lead
Building an AI ecosystem for indigenous Malian languages, centered on speech-first technology and accessibility, with applications across education, healthcare, and business.
| Introduction to MALIBA-AI | Hugging Face Community | Github |
References
Sebastien DIARRA - ML/AI Researcher
sebastien.diarra@gmail.com
Michael Leventhal - Founder & President, RobotsMali
m.leventhal@robotsmali.org
Allahsera Auguste Tapo- NLP Researcher, RIT
aat3261@g.rit.edu
Sidy Ndao Founder & President, Dakar American University of Science and Technology
sndao@daust.org
Abdoulaye Faye - Co-Founder & CTO, CAYTU Robotics
afaye@caytu.com