Benard Adala Wanyande

Bernard Adala Wanyande

Software Engineer

A full-stack developer and MSc Data Science graduate specializing in Generative AI.

Looking for a LeetCode buddy?
Let's connect!

About

I combine strong software development skills with data science expertise to build intelligent, impactful applications, focusing on leveraging AI to create innovative and efficient digital solutions.

Learn More About Me

Experience

NOV 2025 - PRESENT

Teaching Assistant: Quantum Computing ·

Q-Team (THUAS, TU Delft, Leiden University)

Teaching fundamental quantum computing concepts to students through lectures and practical lab sessions. Contributing to curriculum development and course material creation. Organizing educational workshops and events to promote quantum computing education across the collaborative university network.

  • Quantum Computing
  • Project Planning
  • Event Planning
  • Educational Workshops

FEB 2025 - SEP 2025

Data Scientist: Generative AI Specialist ·

Unilever

Automated shelf life prediction using statistical modeling and machine learning, reducing manual analysis time by 60% and increasing forecast accuracy by 30%. Developed an AI-powered learning assistant that summarizes SharePoint content into structured learning paths, cutting onboarding time by 40% and boosting knowledge retrieval efficiency by 50%. Worked on a RAG chatbot on propriety company data using state of the art retrieval models and Open LLMs.

  • Python
  • LangChain
  • OpenAI APIs
  • NextJS
  • FastAPI
  • Azure DevOps
  • Microsoft Copilot

AUG 2024 - JAN 2025

Junior Fullstack Developer ·

Risk Africa Innovatis

Developed Next.js frontend components and FastAPI microservices for the NGCDF project, streamlining fund management. I also integrated OpenAI’s LLM for translations, reducing development time by 75%, and led the migration to a microservices architecture, enhancing system efficiency.

  • Python
  • TypeScript
  • NextJS
  • FastAPI
  • Pandas
  • Agile
  • Jira

View Full Experience

Projects

DutchMaster AI Application

DutchMaster AI

Github Repo Live Site

A multi-model web application for Dutch homework analysis. Uses three AI providers (OpenAI, Anthropic, Google) in parallel to analyze homework images, explain grammar rules, and decode Dutch idioms. Features 10 homework types, 3 model tiers, and PDF export.

  • Next.js
  • FastAPI
  • OpenAI
  • Anthropic
  • Gemini
  • Tailwind CSS
Visualization of Recommender System

Hybrid Pair Based Implicit Recommender

Live Demo: Kaggle Competition Link Github Repo

Developed a hybrid pair-based implicit recommender system for e-commerce, focusing on user interactions and item metadata. Leveraged a classification approach, engineering features for user-item pairs and training multiple machine learning classifiers (Random Forest, LightGBM, etc.). The system provides personalized top-10 recommendations, evaluated based on Recall@10.

  • Python
  • SVD
  • Random Forests Classifier
  • LightGBM
  • Logistic Regression

View Project Archive