Generative AI Architect (Code Generation)

Job Description

Role Summary

We are seeking a highly skilled Generative AI Architect to design and build advanced AI systems for code generation, optimisation, and automation. This role focuses on developing end-to-end intelligent systems leveraging Large Language Models (LLMs), retrieval pipelines, and scalable back-end architectures to enhance developer productivity and modernise software development workflows.

Must-Have Qualifications

Expert in Python with strong proficiency in one or more programming languages; C, C++, Java, and JavaScript; deep understanding of DSA and OOP.

Deep understanding of transformer architectures, tokenization, embeddings, and code-specific LLMs (e.g., StarCoder, CodeLlama)

CodeGen Solutions: Designing, building, and deploying LLM-based applications for code generation, completion, and translation.

Team Leadership: Mentoring junior engineers, defining technical standards, and leading end-to-end development—from POC to production—while designing and delivering AI-powered systems for code generation and developer tooling.

Evaluation & Validation: Developing evaluation suites to ensure the generated code is correct and safe.

Preferred Qualifications

Strong dataset engineering experience: building, cleaning, deduplicating, and structuring large-scale code datasets.

Experience with Hugging Face Transformers, LangChain, LlamaIndex, and vector databases

Strong experience in building RAG pipelines (chunking, retrieval, context optimisation)

Expertise in API design, micro-services, and scalable model deployment

Experience with evaluation systems (test generation, bench-marking, security validation)

Familiarity with Docker, Kubernetes, cloud platforms (AWS/GCP/Azure), and GPU orchestration

Expectations

Long-term thinker with a founder mindset

Ability to translate vision architecture working systems

End-to-end ownership across product, architecture, and execution

Experience in building scalable, production-grade systems

Commitment to building something impactful, not just functional

Strong decision-making under ambiguity and evolving requirements

Bias for action—comfortable working in early-stage, unstructured environments

Benefits

Co-founder position with strategic decision-making authority

Equity allocation up to 5% based on contribution and milestones

Work as a partner, not an employee—full ownership mindset

Opportunity to be listed as co-inventor in patents

Direct influence on product vision, architecture, and roadmap

Opportunity to build and scale a category-defining product (SPROSE)

Flexible work structure with high autonomy and accountability