AI Engineer

Agile Bridge; Pretoria; On-site; ZAR 700,000 - 1,100,000; Full time

Job summary

Agile Bridge is seeking an AI Engineer to turn promising models into secure, scalable AI capabilities. You will design architectures, train and evaluate models, and integrate them into software systems while improving production performance.

Collaborate with data scientists, software engineers and platform teams; focus on production readiness, data quality, latency, security and responsible AI. Expect to build maintainable services, CI/CD pipelines, and monitor AI workloads.

Qualifications

• A bachelor's degree in Computer Science, Artificial Intelligence, Machine Learning, Data Science, Software Engineering, Mathematics or a related quantitative field.
• Typically 3-5 years' relevant experience building and operationalising machine learning or deep learning capabilities.
• Strong Python capability and evidence of sound software-engineering practice.
• Hands-on experience developing, evaluating and improving machine learning or deep learning models.
• Experience integrating model-backed capabilities into a software or production environment.
• The judgement to balance model quality with latency, scalability, reliability, security, maintainability and cost.

Responsibilities

• Translate product and technical requirements into model objectives, architecture decisions and measurable acceptance criteria.
• Select, implement and train machine learning, deep learning, transformer or foundation-model approaches suited to the problem.
• Prepare training and evaluation data, including preprocessing, feature or representation engineering and dataset versioning.
• Run controlled experiments, tune models and evaluate accuracy, robustness, generalisation and computational efficiency.
• Package model-backed capabilities as maintainable services, APIs, libraries or platform components.
• Build or contribute to CI/CD and MLOps pipelines for testing, registration, release and environment promotion.
• Deploy and monitor AI workloads, responding to drift, latency, failures, resource constraints and cost concerns.
• Embed security, responsible AI, traceability and clear technical documentation throughout the model lifecycle.
• Research emerging methods and adopt them only where benchmarking demonstrates meaningful value.

Skills

Python

Software engineering

REST APIs

SQL

PyTorch

TensorFlow

scikit-learn

Git

Docker

CI/CD

Education

Bachelor's degree in Computer Science or related field

Honours or Master's degree (desirable)

Tools

PyTorch

TensorFlow

scikit-learn

Docker

Git

Kubernetes

REST APIs

Job description

Purpose of the Role:

Agile Bridge is looking for an AI Engineer who can turn promising models into secure, scalable and dependable AI capabilities. You will work at the intersection of machine learning, deep learning and software engineering: designing model architectures, training and evaluating models, integrating them into software systems and improving their performance in production.

What to expect in this role

This role is model- and engineering-focused. A project may require you to select and adapt a neural-network architecture, develop a custom prediction or representation capability, benchmark a foundation model, optimise inference performance, or package a validated model behind a reliable service. You will be expected to consider the full production context: data quality, model behaviour, latency, throughput, resource use, cost, security and maintainability.

You will collaborate with Data Scientists on scientific validity and evaluation, Software and Applied AI Engineers on product integration, and Platform teams on deployment and observability. The role is distinct from a purely analytical Data Scientist position and from an Agentic AI role centred on flows, agents and business-process orchestration.

What you will do

• Translate product and technical requirements into model objectives, architecture decisions and measurable acceptance criteria.
• Select, implement and train machine learning, deep learning, transformer or foundation-model approaches suited to the problem.
• Prepare training and evaluation data, including preprocessing, feature or representation engineering and dataset versioning.
• Run controlled experiments, tune models and evaluate accuracy, robustness, generalisation and computational efficiency.
• Package model-backed capabilities as maintainable services, APIs, libraries or platform components.
• Build or contribute to CI/CD and MLOps pipelines for testing, registration, release and environment promotion.
• Deploy and monitor AI workloads, responding to drift, latency, failures, resource constraints and cost concerns.
• Embed security, responsible AI, traceability and clear technical documentation throughout the model lifecycle.
• Research emerging methods and adopt them only where benchmarking demonstrates meaningful value.

What you need

• A bachelor's degree in Computer Science, Artificial Intelligence, Machine Learning, Data Science, Software Engineering, Mathematics or a related quantitative field.
• Typically 3-5 years' relevant experience building and operationalising machine learning or deep learning capabilities.
• Strong Python capability and evidence of sound software-engineering practice.
• Hands-on experience developing, evaluating and improving machine learning or deep learning models.
• Experience integrating model-backed capabilities into a software or production environment.
• The judgement to balance model quality with latency, scalability, reliability, security, maintainability and cost.

Core technical skills

We are looking for depth in the core engineering areas, not superficial exposure to every framework. Comparable tools are welcome where they demonstrate the same capability.

• Programming and software engineering: Strong Python, object-oriented design, testing, debugging, dependency management and clean, maintainable code. SQL is required; C# or Java is useful for integration-heavy environments.
• Machine learning and deep learning: Strong fundamentals across supervised and unsupervised learning, neural networks, loss functions, optimisation, regularisation and generalisation.
• Frameworks: Practical experience with PyTorch, TensorFlow, scikit-learn or equivalent frameworks. Transformer or foundation-model experience should include evaluation and adaptation, not only API consumption.
• Model evaluation and optimisation: Experiment design, meaningful metrics, benchmarking, hyperparameter optimisation, error analysis, robustness testing and performance profiling.
• Data and storage: Data preprocessing, feature or representation engineering, dataset versioning, and experience with relational or non-relational databases.
• Integration: Building or consuming REST APIs, service contracts and model-serving patterns that connect AI capabilities to wider software systems.
• MLOps and delivery: Git, Docker, CI/CD, model versioning or registries, reproducible environments and monitoring of model and service performance.
• Security and governance: Privacy and PII protection, secrets and access management, traceability, responsible-AI evaluation and secure model deployment.

Useful, but not essential

• An Honours or Master's degree in Artificial Intelligence, Machine Learning, Computer Science or a related field.
• Experience with MLflow or an equivalent experiment-tracking and model-registry platform.
• Exposure to Azure, AWS or Google Cloud AI services and infrastructure-as-code practices.
• Experience with Kubernetes, GPU workloads, distributed training or inference optimisation techniques such as batching, quantisation or model compression.
• Practical experience evaluating or adapting generative models, including transformers, embeddings and retrieval components.

How you work

• You think in systems and understand how model behaviour affects the wider product.
• You experiment rigorously and distinguish benchmark improvement from real operational value.
• You communicate technical trade-offs, uncertainty and limitations honestly.
• You collaborate across modelling, software, data and platform disciplines.
• You remain curious without introducing new technology merely for novelty.

The opportunity

You will join an innovation-focused engineering environment where AI capabilities are expected to survive real users, real data and real production constraints. The role offers exposure to model development, production AI, responsible engineering and emerging model architectures while working alongside experienced software, data and platform teams.

How success will be measured

• Model and architecture choices are appropriate to the use case and supported by evidence.
• AI capabilities meet agreed quality, latency, reliability, security and cost requirements.
• Experiments, code, data dependencies and model artefacts are reproducible and maintainable.
• Production performance remains visible, with risks and deterioration addressed early.
• Customers and engineering teams can understand the capability, constraints and operational requirements.