Bunmi Akinremi
Machine Learning Engineer & Researcher | Bridging RL, GenAI, and Scalable Systems
Lagos, Nigeria
Actions
Bunmi Akinremi operates at the convergence of rigorous academic research and industrial-scale engineering. As a Machine Learning Engineer at Kochava, she architects production-grade pipelines that drive decision-making in the advertising technology space. Simultaneously, as an incoming Research Intern at EPFL, she explores the frontiers of Reinforcement Learning (RL) and Generative AI, focusing on how theoretical advancements can solve complex, real-world problems.
Previously, Bunmi engineered the first GenAI-powered fact-checking platform for CJID, leveraging Large Language Models to combat misinformation in journalism. Her technical breadth spans from designing open-source RL platforms to optimizing timeseries data interpolation for cohort analysis.
Beyond the code, Bunmi is a recognised voice in the global AI community. She serves as a technical reviewer for workshops at top-tier conferences, including ICLR and ECCV, and actively builds capacity for the next generation of researchers through Deep Learning Indaba and the Women in Data Science community.
Links
Area of Expertise
Topics
How to pre-record a session
In this session, I'll talk on
- How to set up a place to use for recording (5 mins).
- Tools to use for recording sessions and how they help make recording easier for you (10 mins).
- Tips on how to make a recorded session feel like a live presentation (3 mins).
- Things to avoid while making a recording. (2 mins).
Uncertainty Matters: Why Probabilistic ML is Still Essential in the Age of LLMs
As large language models (LLMs) like ChatGPT continue to dominate AI applications, there is a growing tendency to view them as one-size-fits-all solutions for reasoning, decision-making, and even software development. However, at the core of robust machine learning systems lies a critical concept, Probabilistic Machine Learning. This talk will explore why probabilistic reasoning remains fundamental, even in an era where deep learning and transformer-based models seem to overshadow traditional statistical methods.
We will discuss how probabilistic modelling is a key pillar in machine learning, supporting essential components such as causality, uncertainty estimation, statistical inference, and structured decision-making. The talk will highlight how Probabilistic ML is integral in areas like Bayesian inference, predictive modelling, and reinforcement learning and in improving LLMs' reliability, interpretability, and ability to handle uncertainty. Attendees will learn why fundamental probabilistic principles remain essential for building truly intelligent, explainable, and trustworthy AI despite the rise of LLMs.
LangChains with Azure OpenAI
LangChain revolutionizes your Large Language Model (LLM) capabilities by enabling them to efficiently extract information from APIs, databases, and a myriad of Knowledge Bases, resulting in a seamlessly enriched user experience. With LangChain, you have the power to query your LLM model for real-time internet data.
Join us for an immersive hands-on session where you'll learn how to create and deploy a Language Model on Azure OpenAI and seamlessly connect it with external APIs and databases. We'll delve into advanced techniques for querying databases using natural language, and walk through generated content from an open podcast.
Deploy an Emotion Detection App with Microsoft Azure
Want to learn about building an Emotion detector without having to write ML codes from scratch? I'll show you how to leverage Microsoft Azure AI Services to build a cross-platform solution without having to write a single ML code.
In this session, I'll talk about
- Why emotion detection and what are its use cases?
- Microsoft Azure Vision API.
- A demo walkthrough session where you'll build a simple emotion detection app.
- Useful resources to explore more use cases.
- Finally, a Q&A session to answer all your questions.
Turning research papers into working solutions
Turning a research paper into working codes for the first time can be daunting, difficult, hard, and scary.
For someone who looks into solving complex problems that require research, rather than wasting time and money reinventing the wheel or waiting till someone does it and makes it open source, he can simply use an already existing research paper that provides a solution to his problem. There's the fear of too many paper pages or over-complexity of papers or maths symbols and equations(scary and flags disinterest).
I'll talk about 9 practical steps which I employed to solve this problem which include
- understanding your problem,
- finding papers on that topic,
- filtering papers by skimming through important parts,
- finding relevant data sources and computing resources that will help build your solution,
- creating a roadmap for the methodology,
- supplementing each phase in the roadmap with additional resources to fit your use case,
- comparing your results with the paper’s,
- improving your solution with suggestions included in the summary,
- and reach out to the authors if you have questions or concerns.
By the end of the session, you'll be able to apply the transformation map to any research paper you wish to implement.
The Art of Forecasting
Forecasting is a common word used to mean predicting the future. In business scenarios, you are trying to predict sales, customer activity, or other related metrics, but there's much more to it. Data holds a lot of insight, but data alone is insufficient for decision-making, especially decisions that directly affect revenue, cost, or customers. Other qualitative than quantitative factors affect a business's proposed or predicted outcomes, which must be deliberated alongside the insights obtained from the data.
In this session, we will dive deep into the nitty gritty of forecasting - more art than science, taking a business scenario as a use case and the other physiological and psychological factors that come into play in determining its impact as a decision support system.
DataFestAfrica Sessionize Event
Global Azure 2022 Sessionize Event
Bunmi Akinremi
Machine Learning Engineer & Researcher | Bridging RL, GenAI, and Scalable Systems
Lagos, Nigeria
Links
Actions
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