Daniel Akhabue
AI/ML Engineer
Berlin, Germany
Actions
Daniel Akhabue is an AI/ML Engineer and Cloud Solutions Architect based in Lagos, Nigeria, building production-grade AI systems, with a focus on solving real problems across emerging markets. As an AI Engineer, Daniel architected and built Nigeria's first cloud-native conversational Agentic AI banking solution, which scaled transactional capacity by 70% and cut customer enquiry workloads by two-thirds.
Across industries, EdTech, FinTech, HealthTech, and Supply Chain, Daniel has consistently built with Python at the core, from FastAPI backends and agentic pipelines to PyTorch models deployed on hybrid environments. He is a Kaggle Expert and a four-time hackathon winner, including the Deep Learning IndabaX Nigeria ML Hackathon.
Beyond shipping code, Daniel champions the African AI community. He is a faculty member and organiser at AISOC (AI Summer of Code), a 150+ member community driving hands-on AI learning. He contributes as a researcher with the Machine Learning Collective. He has spoken at PyCon Kenya 2025, Datafest Africa 2024, and Black in Robotics 2025, and served as a community leader at Data Scientists Network (DSN).
Daniel holds a B.Eng. in Computer Engineering from the University of Benin.
He believes the most powerful AI systems are the ones built for the communities that need them most.
Links
Area of Expertise
Topics
Pydantic to the core: Data Validation for Agentic AI Systems.
Pydantic to the core: Data Validation for Agentic AI Systems.
(Discover how Pydantic brings structure, type safety to agentic AI systems.)
The ongoing development of autonomous and agentic AI systems has made data validation an essential requirement. Agentic AI systems frequently operate in unpredictable settings while producing dynamic prompts and making decisions from partially structured data.
With AI software systems becoming more vulnerable to failure because of a single malformed input, this session demonstrates how Pydantic serves as a critical library for Python developers building powerful and type-safe agentic workflows through its advanced data validation capabilities.
The session will teach you to model complex agent data flows while validating them and maintain consistency across various AI toolchains to detect edge cases before they evolve into production problems.
It would also cover different approaches to create dependable AI systems through LangChain agent orchestration, FastAPI pipeline development and custom reasoning loop experimentation.
Techniques to look out for include:
· Data validation (request and response models)
· Data parsing
· Settings management (pydantic.BaseSettings)
· Serialization / Deserialization
· Type enforcement (runtime type checking)
From Students to Leaders: Leveraging Campus Communities for Growth in Data Science
Join us for an electrifying conference session titled "From Students to Leaders: Leveraging Campus Communities for Growth in Data Science" Get ready to dive into the dynamic world of data science and discover the incredible journeys of African students within campus communities.
In this captivating session, we will explore the pivotal roles played by these talented students and how their active engagement propels the growth of campus communities. From driving innovation to fostering collaboration, these students are at the forefront of transforming data science education.
But that's not all! We will also delve into the remarkable ways in which these campus communities serve as catalysts for personal and professional development. Discover the secrets behind leveraging these communities to unleash the untapped potential of aspiring data scientists.
As the session unfolds, we will unveil the remarkable strategies that organizations can employ to give back to these communities and establish fruitful partnerships. By embracing these opportunities, organizations can tap into the incredible talent pool within these campus communities and fuel their own growth.
Be prepared to be inspired and motivated as we showcase the immense power of data science campus communities. Whether you're an aspiring data scientist, an organizational leader, or an education enthusiast, this session is a must-attend. Don't miss out on this transformative experience that will leave you brimming with ideas and ready to unlock new avenues for talent acquisition and collaboration. Join us and be a part of this exhilarating journey towards a brighter future in data science!
COST AT BAY: Building Domain-specific AI Applications using PEFT Techniques
**Join the session with this link:
https://tinyurl.com/3st3p36p
(please copy and paste the above link in a browser)
In an era where technological efficiency and cost management are paramount, the ability to develop domain-specific AI applications that are both effective and resource-efficient is critical. This session, titled "COST AT BAY: Building Domain-specific AI Applications using PEFT Techniques," addresses the pressing business problem of how to create tailored AI solutions without incurring prohibitive computational costs. This is particularly relevant for industries such as healthcare and finance, where the need for specialized AI models is growing rapidly.
This session directly addresses this by showcasing how Parameter-Efficient Fine-Tuning (PEFT) techniques can be utilized to build domain-specific AI applications. By minimizing computational resources while maximizing performance, PEFT offers a practical solution for businesses aiming to leverage AI without the associated high costs.
For data scientists, AI practitioners, and business leaders, the challenge of balancing performance with cost is ever-present. This session is relevant because it provides actionable insights into how PEFT can be applied to create high-performing AI models tailored to specific industry needs. Whether you're working in healthcare, finance, or another data-intensive field, the ability to deploy efficient AI solutions can drive significant innovation and competitive advantage.
In this session, attendees will gain a comprehensive understanding of Parameter-Efficient Fine-Tuning (PEFT) and its advantages over traditional fine-tuning methods. They will acquire practical knowledge on implementing various PEFT techniques, such as adapter layers and Low-Rank Adaptation (LoRA), in their AI development workflows.
Additionally, attendees will learn about the strategic benefits of using PEFT, including reduced costs, faster deployment times, and improved model performance tailored to specific tasks.
Please note that Sessionize is not responsible for the accuracy or validity of the data provided by speakers. If you suspect this profile to be fake or spam, please let us know.
Jump to top