AI Engineer
Dura Digital
WorldwideremotePosted 1 month ago
Skill Required
AI-EngineerMachine-Learning-EngineerBackend-EngineerData-EngineerMLOps-EngineerArtificial-Intelligence-EngineerAI-ML-EngineerAI-Engineering-SpecialistAI-EngineeringArtificial-Intelligence-EngineeringPython-AI-EngineerAI-Engineering-JobsAI EngineerAIsoftware engineeringdata engineeringElasticsearchObservabilityEngineeringJavaScriptTypeScriptsimilar)buildingRAGQuery OptimizationwrittenPythondesignAzureCI/CDCloudDesign PatternsMachine LearningJavaAPIsGenerative AIC#SQLContract
Key highlights
- Required experience: 3+ years of professional software engineering, including 1+ years building with LLMs or ML systems in production.
- Strong Python required, plus working proficiency in at least one of TypeScript/JavaScript, C#, Java, or Go.
- Cloud deployment experience on Azure, AWS, or GCP required.
- Client-facing stakeholder interaction is a core part of the role.
Role overview
Dura Digital is a global digital transformation and consulting company that helps organizations turn emerging technologies into practical business outcomes, bringing together strategy, human-centered design, engineering, and AI expertise to improve operations, customer experiences, and organizational performance. As a Dura Digital consultant, you will represent their commitment to thoughtful innovation, strong client partnership, and high-quality delivery while working closely with the client's leadership and delivery teams.
Responsibilities
- Build and deploy production AI features: LLM-backed workflows, retrieval-augmented generation, agentic and tool-using systems, classification and extraction pipelines, and traditional ML models where they're the better fit.
- Own the data path behind those features: ingestion, chunking, embedding, indexing, and retrieval quality, and improve it based on measured outcomes rather than intuition.
- Design evaluation before you ship. Define what 'good' means for each use case, build offline eval sets and online feedback loops, and track regressions as prompts, models, and data change.
- Integrate AI capabilities into existing applications and services through well-designed APIs, with attention to latency, cost per request, failure modes, and graceful degradation.
- Instrument and monitor deployed systems for quality drift, hallucination rates, token spend, and abuse, and act on what you see.
- Apply responsible-AI practices in day-to-day work: data privacy and residency, PII handling, prompt injection and data exfiltration risks, access control, and clear documentation of model limitations.
- Prototype quickly to reduce uncertainty, then make a deliberate call on what deserves to be hardened and what should be thrown away.
- Contribute to internal standards, reusable components, reference architectures, and patterns other teams can adopt.
- Explain technical tradeoffs to non-technical audiences, including client stakeholders and executives, without overselling what the technology can do.
- Work across the full path from problem framing to deployment: turning an ambiguous business need into a working system, choosing the right model and retrieval approach, building the evaluation harness that proves it works, and standing behind it once real users depend on it.
- Partner closely with product, design, data, and platform engineers, and depending on the engagement, directly with client stakeholders who are new to AI and need a clear-eyed guide on what's realistic.
Requirements
- 3+ years of professional software engineering experience, including 1+ years building with LLMs or ML systems in production.
- Strong Python. Working proficiency in at least one of TypeScript/JavaScript, C#, Java, or Go for service and application integration.
- Demonstrated experience shipping an AI feature end to end; you can walk us through what you built, how you evaluated it, what broke, and what you changed.
- Practical command of the modern LLM toolkit: prompt design, structured output, function/tool calling, context management, and at least one orchestration framework or a well-reasoned argument for building without one.
- Experience with retrieval systems and vector search (e.g., pgvector, Azure AI Search, Pinecone, Elasticsearch, or similar), including how to diagnose bad retrieval.
- Cloud deployment experience on [Azure / AWS / GCP], plus containers, CI/CD, and infrastructure-as-code fundamentals.
- Solid data engineering instincts: SQL, working with messy real-world data, and building pipelines that don't silently fail.
- Clear written and verbal communication, and comfort operating with incomplete requirements.
Nice to have
- Consulting, agency, or client-facing delivery experience.
- Fine-tuning, distillation, or parameter-efficient adaptation of open-weight models, and a sense for when it's worth it versus prompting or retrieval.
- Experience with agent frameworks, MCP, or multi-step tool-using systems in production.
- MLOps tooling (experiment tracking, model registries, feature stores) and observability platforms for LLM applications.
- Background in a regulated domain such as healthcare, financial services, insurance, or the public sector.
- Speech, vision, or document-understanding work (OCR, layout parsing, multimodal models).
- Open-source contributions, technical writing, or conference speaking.
Additional details
- This is a hands-on engineering role, not a research role.
- As a Dura Digital consultant, you will represent our commitment to thoughtful innovation, strong client partnership, and high-quality delivery while working closely with the client's leadership and delivery teams.
- Originally posted on Himalayas.