DigiTribe - Machine Learning Engineer

Your role & responsibilities

As a Machine Learning Engineer, you will :

Design, develop, and deploy ML and NLP models to solve business problems in areas like customer service, compliance, fraud, and operational efficiency.

Build and optimize retrieval-augmented generation (RAG) pipelines integrating LLMs and knowledge bases.

Develop and enhance chatbots and virtual assistants using GenAI and conversational AI frameworks.

Collaborate with cross-functional teams to deliver AI products on Microsoft Azure cloud infrastructure.

Automate end-to-end ML workflows using tools like Azure ML, Databricks, and MLOps pipelines.

Ensure models are secure, ethical, explainable, and compliant with financial industry regulations.

Contribute to continuous learning, experimentation, and evaluation of new GenAI tools, open-source frameworks, and model architectures.

Key qualifications

Must-haves

3+ years of experience in Machine Learning, with a focus on NLP or AI applications

Strong Python programming skills (e.g., scikit-learn, PyTorch, Transformers, Langchain)

Solid experience working with Azure cloud services (Azure ML, Azure Functions, Azure DevOps, etc.)

Hands-on experience with LLMs, RAG frameworks, vector databases, and embedding models

Background in building or scaling chatbots, Q&A systems, or conversational AI

Familiarity with data engineering principles and MLOps practices

Good understanding of model evaluation, bias detection, and explainability in ML systems

Nice-to-have

Experience with tools like Openai API, Hugging Face, Langchain, Weaviate, Pinecone, or Qdrant

Knowledge of regulatory or compliance frameworks in banking (e.g., GDPR, data privacy)

Previous experience in the financial services or banking industry

Soft skills

Strong analytical and problem-solving skills

Clear and effective communicator, both with technical teams and business stakeholders

Passion for experimentation and emerging AI trends

A team player with a proactive and solution-driven mindset