Data Engineer - AI, NLP, Vector Search

January 23

Apply Now

Receive Emails with Similar Jobs

Kupa Global

WebsiteLinkedInAll Job Openings

Kupa Global is a talent management and recruitment company that specializes in building and enhancing cross-border teams, particularly in South Africa and Kenya. They provide services such as talent sourcing, payroll management, and offering office space. With a presence in London, Cape Town, and Nairobi, Kupa Global helps companies integrate top problem-solvers from Africa into their teams. The company also provides comprehensive support, including mentoring, coaching, and light-touch oversight, to ensure team productivity. Their candidate selection process is based on a proprietary potential-based hiring model and they pride themselves on a high success rate and low churn rates for their placements.

Talent Acquisition • Talent Advisory • Remote work

2 - 10 employees

🎯 Recruitment

🤝 B2B

👥 HR Tech

📋 Description

About RVS • Our client, Royal Voluntary Service (RVS), is a leading UK charity managing one of the largest networks of volunteers across the country to deliver essential healthcare, homelessness, and food security services. • RVS is embarking on an exciting digital transformation and hiring their first-ever tech team out of South Africa in their 80-year history. • We are seeking a mid-level DevOps Engineer to join RVS’s technology team. • You’ll play a critical role in developing stable, scalable platforms that empower volunteers to continue their life-changing work. • You will help enforce best practices in coding, automation, and infrastructure management, thus driving faster, more reliable software releases. • This role reports to the IT Service Delivery Manager. • Consider applying if you want to make an impact whilst earning a competitive salary. What You Will Be Doing • Vector Search Solutions: Design and implement vector-based search systems (e.g., ChromaDB) and optimize performance for large-scale datasets, supporting both real-time and batch queries. • LLM Integration: Install, fine-tune, and deploy large language models like Llama 2 (Note: using ChatGPT will not suffice!) and develop workflows for generating high-quality text summarizations and embeddings. • Model Training & Fine-Tuning: Train and adapt LLMs using domain-specific datasets, continuously evaluating and improving model accuracy, scalability, and efficiency. • Data Engineering & Delivery: Develop and maintain robust ETL pipelines in Python, use Docker for containerization, and implement CI/CD pipelines to streamline integration and delivery. • Documentation & Best Practices: Thoroughly document workflows, codebases, and best practices to ensure long-term maintainability and scalability.

🎯 Requirements

Our Ideal Candidate Has : • Experience & Knowledge: • 5+ years in Data Engineering roles with a strong background in Python (Pandas, NumPy, PyTorch). • Proven track record working with LLMs (e.g., Llama 2) and vector databases (e.g., ChromaDB). • Familiarity with containerization (Docker) and CI/CD pipelines (e.g., Jenkins, GitHub Actions). • Technical Skills: • Skilled in setting up AI/ML workflows in cloud environments (AWS, GCP, or Azure). • Experience with distributed computing frameworks (Spark, Dask) and additional vector search systems (Milvus, Pinecone) is a plus. • Comfortable integrating RESTful APIs, fine-tuning models, and optimizing performance at scale. • Problem-Solving & Communication: Strong analytical and troubleshooting abilities. An effective communicator able to collaborate across multidisciplinary teams and explain complex concepts clearly.

🏖️ Benefits

In addition to a very competitive salary we have additional perks including: • A healthcare stipend in addition to a competitive monthly salary • Opportunity to contribute to a meaningful cause and see the direct impact of your work. • Flexible hybrid working options for a better work-life balance. • Room for professional growth and skill development through ongoing training and support. • Collaborative and inclusive team culture that values everyone’s input.