Position: GenAI Python Developer

Location: London (and client sites some within an hour of London)

Working Pattern: Hybrid (3 days per week in office or client sites)

The Company

Sandhata is a global consultancy providing Next Gen Engineering Services dedicated to transforming businesses. Our services span all facets of Digital Transformation, DevOps & Cloud, Integration, Automation, Low Code Application Development and AI-enabled services. Sandhata is a privately held IT services company with bases in the UK and India operating worldwide to help clients deliver a 'digital first' strategy.
Sandhata is currently working with clients across various industry sectors, helping them to deliver cutting-edge technology solutions that streamline business processes and drive IT efficiency.

Overview

We are seeking an experienced GenAI Developer to develop core services, features and capabilities on an enterprise-grade Generative AI Virtual Agent front line capability for a large global insurer. This role blends strong Python engineering with DevOps enablement and hands-on integration of LLM technologies into secure, scalable enterprise environments.

The ideal candidate will bring practical development and platform engineering experience, strong DevSecOps skills, and the curiosity to solve emerging engineering challenges in this rapidly evolving Gen AI space.

Key Responsibilities

Familiarity with GenAI concepts such as LLMs, RAG patterns, prompt-driven workflows, and AI service orchestration
Build and maintain Back End services and APIs enabling secure access to LLM capabilities.
Develop and optimise Python-based GenAI components including prompt orchestration, output validation, and evaluation tooling.
Integrate LLMs with enterprise systems, observability, and security frameworks.
Design and maintain CI/CD pipelines aligned to engineering standards (Azure DevOps primarily).
Collaborate closely with platform leads, architects, and design teams to ensure reliable operationalisation of GenAI services.
Support benchmarking, evaluation, and experiment tracking for LLM performance and cost.
Contribute to RAG implementations and data access patterns supporting enterprise use cases.
Help shape API standards, reusable patterns, and documentation for platform adoption.
Troubleshoot and optimise performance across distributed systems and cloud services.

Mandatory Skills & Experience

Strong Python and FastAPI engineering in production systems.
Working knowledge of GenAI technologies and Large Language Models.
Experience evaluating LLM performance and prompt handling complexities.
Solid DevOps mindset with CI/CD expertise and observability best practices (Azure DevOps preferred).
Comfortable working in regulated enterprise environments with strict security controls.
Experience integrating AI services into real-world apps or workflows.
Infrastructure-as-code (Terraform)
Knowledge of authentication, secret management, network boundaries, and model access governance.
Developement of AWS services: EC2, EKS, S3, SQS, DynamoDB, Bedrock.

Preferred Skills

Kong API Gateway, Kong Mesh, Flux CD.
RESTful API development (FastAPI preferred), microservices, Terraform with GitOps workflows.
Prompt evaluation, observability & red teaming tools (like Arize, Promptfoo).
Experience with SQL (MySQL, PostgreSQL) and NoSQL databases
Exposure to RAG patterns and vector search technologies.

Behavioural Competencies

Proactive self-starter who identifies gaps and drives solutions without waiting for tickets.
Comfortable working in ambiguous emerging domains where best practice is evolving.
Collaborative communicator able to influence and gain trust across engineering, data, and product teams.
Curiosity to explore and adopt new GenAI techniques in a structured, secure way.
Bias toward automating everything to reduce toil and accelerate delivery.

What Success Looks Like

Efficient, secure, and reusable GenAI components delivered into production.
Improved engineering velocity through automation and DevOps best practices.
Stable services, with observability and evaluation built in from day one.
Platform teams and application teams love using what you build.
Clear documentation and fast onboarding for future use cases.