AI Product Developer

About Audens AIAudens AI is an enterprise AI communications solution provider for Marketing, Sales, Corporate Communications, Media, Reputation and Brand teams. We help companies improve how they are discovered, understood and represented by AI systems, and then build AI-native products that turn enterprise knowledge, data and communications intelligence into action.Our work spans two connected layers:UPSTREAM: AI VISIBILITYMaking organisations, brands, products, evidence and expertise more visible and usable across LLMs and AI search—including AI visibility diagnostics, source intelligence, GEO/AEO/LLMO, structured knowledge, content architecture, retrieval readiness and authoritative third-party signals.DOWNSTREAM: AI APPLICATIONSBuilding agents, agentic workflows, orchestration layers, RAG applications, simulation-based cohorts, synthetic audiences, digital twins, decision-support systems and other AI products for enterprise communications and commercial use cases.Website Link: https://audens.ai/About the RoleWe are looking for a hands-on AI Product Developer who can build production-grade AI systems in code and help shape the Audens AI product stack. This is not a generic chatbot role. You will work across the full AI communications lifecycle — from systems that measure and improve AI visibility to agents, orchestration, simulation-based cohorts and digital twins that help companies understand audiences, test decisions and execute work.You should be a strong software engineer with a product mindset: comfortable moving from an ambiguous client or business problem to architecture, prototype, evaluation and production deployment. Python will be the primary development language, with JavaScript/TypeScript useful for product interfaces and integrations.What You’ll Build· AI visibility and knowledge systems that assess how brands, companies, executives, products or issues appear across AI answers, identify source and evidence gaps, and recommend or enable remediation.· RAG and retrieval pipelines that make enterprise content, evidence and approved knowledge usable by AI applications with strong grounding, provenance and access control.· Agents and agentic workflows for research, content intelligence, media/reputation monitoring, brand analysis, sales enablement, corporate communications and other enterprise use cases.· Orchestration layers that coordinate multiple models, tools, data sources, agents and business rules across multi-step workflows.· Simulation-based cohorts and synthetic audiences that allow teams to test messages, propositions, scenarios and likely stakeholder reactions before real-world deployment.· Digital twins and behavioural models representing customer, stakeholder, media or other defined cohorts where the underlying evidence and methodology can support meaningful simulation.· Reusable platform components that can be configured across multiple Audens AI clients instead of being rebuilt account by account.Core ResponsibilitiesProduct & Engineering· Design, build and ship AI-powered applications primarily in Python, using appropriate LLM and agent frameworks such as LangGraph, LangChain, AutoGen, CrewAI or equivalent tools.· Build multi-step agentic systems with tool use, memory, planning, structured outputs, function calling and human-in-the-loop controls where required.· Implement advanced RAG techniques including hybrid search, re-ranking, contextual chunking, metadata filtering, long-context strategies and grounding checks.· Integrate OpenAI, Anthropic, Google and open-source models with enterprise systems, APIs, databases and knowledge sources using standard APIs, function calling and MCP-style integrations.· Create model- and vendor-flexible architectures so products can switch or combine models based on task quality, latency, cost, privacy and client requirements.· Build production-grade APIs, services, data pipelines and connectors; deploy and operate workloads on Azure, AWS or GCP.· Design for enterprise scale, including performance, throughput, latency, concurrency, caching, asynchronous processing, cost controls, observability and reliability.Evaluation, Safety & Reliability· Create evaluation frameworks for accuracy, retrieval quality, hallucination, task completion, latency, cost and regression testing.· Build golden sets, automated judging, human evaluation loops and monitoring appropriate to the product and use case.· Apply security-by-design, data minimisation, access control, auditability, prompt-injection defence, content safeguards and enterprise guardrails.· Clearly distinguish outputs based on sourced evidence from simulations, inferred behaviour or synthetic data; design products so users can understand those boundaries.Client & Product Translation· Work with Audens AI strategy, communications and account teams to translate client problems into technical product specifications and prototypes.·  Rapidly prototype new concepts and validate them with internal teams or real users before hardening them for production.· Support pitches, workshops and client presentations with technical demos, solution architecture and feasibility input.· Help convert recurring client problems into scalable Audens AI products, modules, APIs and reusable workflows.· Stay current with developments across LLMs, agents, model context protocols, retrieval, simulation, synthetic data, digital twins, evaluation and observability — and bring relevant capabilities into the Audens AI stack.Required Qualifications· 3+ years in software or product development, with 1–2+ years of hands-on experience building LLM-powered applications, RAG systems or AI agents.· Strong proficiency in Python; working proficiency in JavaScript/TypeScript is an advantage.· Demonstrated experience taking an AI product or system beyond a demo into a robust prototype or production environment.· Hands-on experience with LLM APIs, prompt and context engineering, structured outputs, tool/function calling and agent orchestration.· Strong understanding of RAG and retrieval systems, including vector databases, hybrid retrieval, re-ranking and grounding strategies.·  Experience designing and integrating APIs, webhooks and enterprise data/connectors; familiarity with MCP is strongly preferred.· Working knowledge of at least one major cloud platform—Azure preferred, with AWS or GCP also relevant — including deployment, monitoring and operations.· Strong software-engineering fundamentals: clean APIs, testing, debugging, observability, performance tuning, CI/CD and secure-by-default design.· Ability to own ambiguous problems end-to-end and drive them toward measurable business outcomes.· Comfort working with client-facing teams and explaining complex technical ideas in simple business language.Preferred / High-Value Experience· Experience building multi-agent systems, agent-to-agent communication or orchestration across models and tools.· Experience with AI evaluation and observability tooling such as LangSmith, Ragas, Promptflow, Azure Monitor or equivalent platforms.· Experience with synthetic data, simulations, persona/cohort modelling, behavioural modelling or digital-twin architectures.· Experience building analytics or decision-support products for marketing, sales, customer experience, communications, media intelligence, reputation or brand teams.· Familiarity with Microsoft 365, Dataverse, Copilot Studio or Power Platform, where these are appropriate to enterprise deployment.·  Experience with workflow/automation platforms such as n8n, Make or Zapier, while retaining the ability to build core production systems in code.· Experience in an AI startup, product company, agency or fast-paced client-services environment.· A strong GitHub, portfolio or demonstrable set of AI/agentic projects.What Makes You a FitYou are an engineer who thinks in products rather than isolated demos. You are comfortable living in code, reasoning about architecture and data flows, and making pragmatic trade-offs among model quality, latency, cost, security and user experience.You also understand that Audens AI sits at the intersection of technology and communications. The objective is not simply to make an LLM perform a task; it is to build systems that help enterprises become more discoverable to AI, use their knowledge more intelligently, understand audiences and stakeholders better, and automate or augment high-value communications and commercial workflows.You move quickly, but you also know when reliability, evidence, provenance and governance matter more than speed. You can build a rough proof-of-concept fast—and then turn the right ideas into products that clients can trust.