Senior AI LLM Engineer

e-Merge IT Recruitment; Sandton; On-site; ZAR 1,020,000 - 1,380,000; Full time

Job summary

e-Merge IT Recruitment seeks a Senior AI Engineer for a permanent role based in Sandton, SA. You’ll build real-time AI solutions across LLMs, voice agents, automated testing, and modern software practices.

With 5+ years engineering and 2+ years in production LLM/speech AI, you’ll own multi-system features and contribute to AI tooling, cloud services, and scalable architectures in a hands-on environment.

Qualifications

• 5+ years engineering experience.
• 2+ years applied LLM / speech AI in production.
• Strong Python experience (APIs, typing/validation, testing).
• OpenAI / Azure OpenAI experience for chat and structured outputs.
• Azure messaging, storage, and SQL-backed services.
• Proven ownership of ambiguous, multi-system features.

Responsibilities

• Build real-time voice agents with modular prompts and multi‑assistant flows.
• Improve automated QA pipelines including VAD and diarization.
• Implement structured LLM pipelines with schemas and batching.
• Own features in Azure Functions / Kubernetes workers with queues and SQL integrations.
• Contribute to post-call intelligence features and evaluation suites.

Skills

Python

LLM / speech AI

OpenAI / Azure OpenAI

ambiguous multi-system features

ownership of features

Tools

Cursor

Claude

AI coding assistants

Kubernetes

SQLAlchemy / Azure SQL

Azure Functions

Job description

Build what’s next in AI.

Join an innovative technolog y environment focused on developing intelligent, AI-powered conversational solutions that are shaping the future of how people interact with technology.

This is a hands‑on opportunity to work across LLMs, voice agents, automated testing, AI coding tools, and modern software development practices. You’ll take on complex technical challenges, contribute to solution design, and help turn ambitious AI concepts into scalable, production-ready solutions.

For a Senior AI Engineer who enjoys solving challenging problems, working with emerging technology, and having a meaningful impact on technical direction, this is an opportunity to step into a role where your expertise can make a real difference.

Must-have skills and experience:

5+ years engineering; 2+ years applied LLM / speech AI in productionStrong Python (APIs, typing/validation, testing, packaging)OpenAI / Azure OpenAI (chat, structured outputs, batch patterns)Azure messaging, storage, and SQL-backed servicesProven ownership of ambiguous, multi-system features

Strong advantages:

Experience with Cursor, Claude, or similar AI coding assistantsVoice agent stacks; latency, barge-in, end pointing awarenessSpeech ML in production (PyTorch, Whisper, diarization, ONNX, GPU Docker)vLLM / guided decoding; Kubernetes worker patternsContact-centre QA or collections domainSQLAlchemy / Azure SQL; IaC awareness or strong infra collaboration

Responsibilities:

Build and harden real-time voice agents: modular prompts, multi-assistant flows, tool-calling, webhooks, telephony/operational concernsImprove automated QA pipelines: VAD, diarization, language gating, transcription, alignment/attribution, LLM scorecard scoringImplement structured LLM pipelines (schemas, batching, failure handling, score recalculation)Own features in Azure Functions / Kubernetes workers: queue consumers, artifact pipelines, SQL integrationsContribute to post-call intelligence features such as outcome scoring, summary generation, and structured data extractionExpand and maintain evaluation suites (scenarios, LLM-as-judge, CI regression)Contribute to email/WhatsApp agent work with reuse of tools, outcomes, and observability

Reference number for this position is GZ61523 which is a permanent position based in Sandton offering a salary of R1.2m per annum, negotiable based on experience and ability.