r
ridwan_adeba

Ridwan A

@ridwan_adeba

ML Engineer for Self Hosted LLM and RAG Systems

Nigeria
Anglais, Russe, Yoruba
Certaines informations sont présentées en anglais.
À propos de moi
I build LLM and retrieval systems for organisations with strict privacy constraints and compliance perimeters. I specialize in shipping production NLP pipelines that run entirely on-premises with zero external API calls, ensuring every output is traceable and auditable. My expertise includes RAG pipelines, fine-tuning LLaMA models, and deploying scalable MLOps solutions.... Plus d’infos

Compétences

r
ridwan_adeba
Ridwan A
hors ligne • 

Voir mes services

Modèles génératifs
I will build ai agents and automation workflows with langgraph and python
Mise en œuvre et déploiement de l’IA
I will build a private self hosted rag chatbot over your own documents

Expérience professionnelle

Sberbank

Data Analyst

Sberbank • Temps plein

May 2026 - Aug 2026 • 3 mos

Analysed ~87,000 software-release comments across the bank’s top five business blocks to rank the root causes of test failures. Built a two-stage NLP pipeline pairing a locally hosted LLM (Qwen) with a transparent rule engine, reaching 96–99% agreement with hand-labelled ground truth. Ran entirely on-premises with zero external API calls.

VK

Machine Learning Intern

VK

Jun 2025 - Aug 2025 • 2 mos

Built a gender-prediction model for ad personalisation end to end: feature engineering from 750,000+ user behaviour records, through training (Random Forest, 81.7% accuracy, 0.887 AUC), to production scoring of 75,000+ users.

VarsityScape

Machine Learning Engineer

VarsityScape • Temps plein

Oct 2022 - May 2024 • 1 yr 7 mos

Led development of ScapeAI, an LLM-powered course-creation platform built on RAG pipelines. Deployed models on Google Cloud behind FastAPI REST endpoints, documented for the frontend and backend engineers integrating against them, and built the production pipeline with automated drift monitoring. Worked with prospective customers from first call to close: discovery, demos, scoping, and onboarding new clients onto the platform once they signed.