LLM Engineer (AI Agents)

Location: Remote
Employment Type: Full-time
Experience Level: Senior / Experienced

About Us

At Wisy, we’re building intelligent AI agents that convert fragmented retail data into real-time decisions and smart actions. By combining large language models, contextual retrieval systems, and cutting-edge agentic workflows, we help global CPG brands gain the clarity and agility they need to win at the shelf. If you’re passionate about pushing the boundaries of AI in production, we want you on our team.

About the Role

As an LLM Engineer, you’ll design, implement, and optimize custom agentic systems that power our platform’s intelligence layer. This role blends applied research and production engineering — ideal for someone excited about taking language models beyond prototypes into scalable, real-world deployments that solve business-critical challenges in the CPG and retail space.

Core Responsibilities (What You’ll Do)

• Integrate LLM capabilities into platform workflows using structured outputs and tool-calling mechanisms.
• Design and implement LLM-based solutions for real-time decision-making and retail intelligence.
• Develop agentic architectures and orchestrate multi-step LLM workflows
• Apply parameter-efficient fine-tuning methods (e.g., LoRA, prefix tuning) to adapt LLMs to domain-specific tasks and data.  
• Collaborate on prompt engineering strategies to maximize model performance across contexts.
• Build human-in-the-loop pipelines that balance automation with user control, enabling oversight and feedback in AI-driven workflows. 
• Evaluate model outputs using internal benchmarks and modern eval frameworks (e.g., DeepEval or similar), ensuring quality and alignment with business needs.
• Contribute to experimentation with both proprietary and open-source LLMs across performance, latency, and cost tradeoffs.
• Work with cross-functional teams to design productized LLM applications at scale.
• Participate in rapid prototyping, iteration, and internal testing of new LLM-powered features.
• Write clean, testable code and maintain reusable modules for prompt, inference, and evaluation pipelines.
• Maintain awareness of trends in agentic LLMs, vector databases, retrieval-augmented generation (RAG), and emerging tooling.
• Document and share insights on model performance, failure modes, and deployment considerations.
• Contribute to internal tooling for debugging, tracing, and analyzing LLM-based workflows.
• Support the development of product and research roadmaps by exploring new LLM capabilities.
• Use AI-assisted development and model-debugging tools to streamline prompt iteration, model evaluation, and performance tuning.

What We Value

• Curiosity-driven problem solving.
• High standards of quality and reproducibility.
• A “builder” mindset — from prototype to production.
• Deep ownership and accountability.
• Strong collaboration across disciplines (product, design, infra).
• Enthusiasm for AI safety, explainability, and ethical deployment.
• A continuous learning and research-driven approach.

What We Require (What We’re Looking For)

• 3+ years of experience working with machine learning, NLP, or AI systems.
• Proficiency with Python and experience building applications with LLM integration
• Familiarity with LLM frameworks such as LangChain, Pydantic AI, or similar.
• Hands-on experience with LLM platforms and model providers, including OpenAI, Hugging Face, and Google/Gemini.
• Understanding of infrastructure tools such as vector databases (e.g., Pinecone, Weaviate, FAISS).
• Comfort with prompt engineering, evaluation frameworks, and model experimentation.
• Experience working with structured data, APIs, and production environments.
• Ability to balance tradeoffs between accuracy, latency, and cost in model deployment.
• Strong communication skills and the ability to explain technical work clearly.
• Startup mindset: bias for action, high ownership, and adaptability