Offiong Bassey Edet
Machine Learning Researcher — Speech & language technology for low-resource African languages
Hi, I'm Offiong Bassey, a machine learning engineer and researcher specialising in natural language processing, with work spanning audio processing, speech synthesis, machine translation, and real-time conversational AI. My research centres on low-resource and multilingual language settings, particularly Efik, an extremely low-resource tonal language spoken in southeastern Nigeria.
I currently work as a Machine Learning Research Engineer at Plotweaver, where I build multilingual speech systems for African languages, and I founded EfikAI, a research initiative building the first publicly documented machine translation, speech recognition, and text-to-speech systems for Efik. I'm also a volunteer researcher with Howard University's NLP Group, a Machine Learning Researcher with Machine Learning Collective, and a research community member of Masakhane.
#1
Research Interests
- Natural Language Processing
- Speech Recognition & Synthesis
- Machine Translation
- Low-Resource Languages
- African Language Technology
- Conversational AI
- Scaling Laws
- Emotional Speech, Paralinguistics
#2
Selected Publications
#3
Research Experience
- Conduct research on LLM-based question answering, multilingual NLP, speech processing, and speech synthesis, with a focus on low-resource and domain-specific settings.
- Clean and preprocess extremely low-resource language data for multilingual NLP tasks, including question answering, machine translation, audio processing, and speech synthesis.
- Conduct research on multilingual NLP and speech AI, with a focus on low-resource language modeling.
- Lead research on Efik language technologies, including machine translation, ASR, and TTS.
- Member of the pan-African NLP research community advancing language technology for African languages.
- Participate in community research discussions, workshops, and knowledge-sharing on multilingual NLP.
- Founded EfikAI, building the first publicly documented MT, ASR, and TTS research systems for Efik.
- Curated 13,865 parallel English–Efik sentence pairs with linguists and native speakers.
- Developed natural-sounding ASR and TTS systems using Whisper, XLS-R, VITS, and diffusion-based models.
#4
News
June 2026
Paper accepted to Interspeech 2026, Sydney, Australia.
March 2026
Paper, Adapting Foundational ASR Models to Efik, accepted to Speakable @ LREC 2026, Spain.
February 2026
Awarded full sponsorship to attend the AfricaNLP Workshop @ EACL 2026, Rabat, Morocco.
February 2026
Joined Plotweaver as a Machine Learning Research Engineer.
January 2026
Gave a talk at Kabod Group on funding opportunities for African language industry projects.
January 2026
First paper accepted to AfricaNLP @ EACL 2026, Morocco.
#5
Positions
Jan 2026 — Present
Machine Learning Engineer — NLP
Multilingual text-to-speech and real-time ASR → MT → TTS pipelines for Yoruba, Swahili, Hausa, and other African languages; reduced end-to-end latency by 87%.
Jun 2025 — Oct 2025
AI Engineer
Conversational AI systems and Rasa NLU chatbots for fintech customer support automation.
May 2024 — Present
Full Stack Engineer
Real-time transcription pipelines and LLM-based summarisation for an AI note-taker used in live classes and meetings.
Jun 2023 — Oct 2023
Backend Developer Intern
RESTful APIs in Node.js and database performance tuning.