AI Engineer

CommVersion is looking for a hands-on AI Engineer to build, enhance and experiment with Generative AI and Agentic AI applications.
You will work across AI Assist, Agentic RAG, conversational AI, summarization, NER, AI-powered message generation, automated responses and other emerging AI capabilities.
The role involves working across the AI application lifecycle from rapid POCs and experimentation to development, testing, evaluation, deployment and continuous improvement.
We're looking for someone who enjoys solving ambiguous problems, experimenting with new AI technologies and turning successful ideas into reliable applications.

Who we are?
Our mission is to maximize the conversion potential of EVERY website, turning live chat into an
indispensable revenue solution for businesses across the world. And we do this through our
SmartChat solution that has transformed live chat from a customer support cost to a conversion
rate optimisation tool for over 500 global brands, supercharging their website performance in
the process.
In ten years, we have grown from a founder-led business to a team of over 150+ employees
spanning four countries.
However, we are still at the tip of the iceberg. So, if you re looking to join a fast-paced ambitious company with a customer centric mindset, then look no further.

Build and enhance LLM-powered applications, AI agents and RAG pipelines.Design AI workflows involving LLMs, tools, APIs, retrieval and multi-step workflows.
Work with LLMs such as OpenAI, Claude and Gemini.
Build and improve conversational AI capabilities including summarization, NER, message generation and automated responses.

Work with vector databases, embeddings and retrieval pipelines to build knowledge- grounded AI applications.
Develop POCs and experiments to evaluate new models, frameworks and approaches.

Implement AI evaluation, testing and guardrails to improve the reliability and quality of AI applications. Work with AI observability to monitor traces, latency, token usage, cost and application
performance.

Integrate AI agents, tool calling and external APIs/services into AI applications. Understand and experiment with emerging AI approaches such as multi-agent systems
and agentic architectures.
Deploy AI applications to Dev and Pre-Prod environments, working closely with the
DevOps team.
Investigate and improve AI application quality, latency, reliability and cost.
Collaborate with engineering, data, product, and other teams to turn business
requirements into practical AI solutions.
What We're Looking For:
Must Have
3+ years of software/AI engineering experience, with practical experience building
AI/GenAI applications
Strong Python programming and backend/API development skills.
Hands-on experience building LLM-powered applications.
Strong understanding of RAG, embeddings, vector databases, and vector search/retrieval.
Practical experience with AI agents, tool calling, and agentic workflows.
Hands-on experience with at least one AI/LLM orchestration framework such as LangGraph, LangChain, LlamaIndex, or equivalent.
Understanding of AI evaluation, including designing test cases and evaluating the quality and reliability of LLM outputs.
Understanding of AI observability concepts, including traces, latency, token usage, cost, and model/application performance.
Understanding of guardrails, hallucination mitigation, and reliable AI application design.
Strong understanding of prompt engineering, including system prompts, few-shot
prompting, structured outputs, context management, and iterative prompt optimization.
Strong software engineering, debugging, and problem-solving skills.
Experience writing unit tests and working with production-quality code.
Basic understanding of Docker and application deployment.
Ability to work with APIs, third-party services, and external tools.

Good to Have
Hands-on experience with AI observability platforms such as Langfuse, LangSmith, or equivalent.
Experience with automated prompt optimization and prompt-tuning frameworks such as GEPA, DSPy optimizers, or equivalent approaches.
Experience with multiple vector databases such as Pinecone, FAISS, Chroma, Weaviate, Milvus, or pgvector.
Exposure to multi-agent architectures and emerging AI protocols such as MCP and A2A.
Experience with AI evaluation frameworks such as Ragas, DeepEval or equivalent.
Experience with AI API gateways and model infrastructure such as LiteLLM, Portkey, Kong
AI Gateway or equivalent.
Experience with web retrieval/search technologies such as Tavily, Firecrawl, or equivalent.

Experience with model routing, fallback strategies, caching, and AI cost optimization.
Exposure to ML/NLP/Deep Learning, including Transformers, Hugging Face, PyTorch, or model fine-tuning.

Experience with Kubernetes and CI/CD.
Experience working with multiple LLM providers and models.
What We're Looking For in You
We're looking for someone who can understand an existing AI system quickly, take ownership
of features, experiment rapidly, and turn successful POCs into reliable production systems.

You should be comfortable working across the LLM application stack from prompting,
orchestration and retrieval to testing, evaluation, observability and deployment.
You don't need to know every tool in our stack. Strong AI engineering fundamentals, practical
experience, curiosity and the ability to learn quickly matter more than a long list of frameworks.
We're particularly interested in engineers who can take an ambiguous AI problem, break it
down, make sensible technical decisions, build a working solution, and continuously improve it based on real-world results.