LLM Expert

At ZLAB, we work in a dynamic and collaborative environment where creativity and innovation thrive. Our team is passionate about leveraging intelligent solutions to tackle real-world challenges and push the boundaries of what’s possible. We value teamwork, curiosity, and growth, creating a space where everyone contributes to building smarter, future-ready technologies.

About the project

hAIre is an AI-powered hiring support system designed to improve efficiency, consistency, and fairness across the recruitment process. The system combines structured data analysis with LLM-driven reasoning to assess candidates, detect risks, and enhance hiring decisions.

Role overview

We are seeking an LLM Expert to design, implement, and optimize large language model pipelines that power the reasoning and decision-making capabilities within hAIre. You will work on fine-tuning, prompting, evaluation, and optimization of LLM-based systems across multiple modules. Your goal is to make the models accurate, efficient, explainable, and aligned with HR best practices.

Key Responsibilities

-   Design, implement, and test LLM-based components

-   Develop prompt frameworks and scoring logic for text data

-   Optimize inference flows by balancing structured and black-box LLM evaluations.

-   Design evaluation benchmarks for model accuracy, bias detection, and response quality.

-   Collaborate with backend engineers to integrate model outputs with APIs built on FastAPI, Pydantic, and Dockerized microservices.

-   Prototype and test fine-tuned or instruction-adapted models (e.g., GPT-4o, Llama-3, Mistral) to reduce cost and latency.

-   Contribute to bias mitigation logic through careful prompt design and alignment techniques.

-   Define evaluation datasets using real and synthetic datasets

-   Monitor production model performance and iteratively refine prompt templates and fallback logic.

Requirements

-   Proven experience working with large language models (e.g., GPT, Claude, Gemini, Llama-3) in applied contexts.

-   Proficiency with Python, LangChain, OpenAI SDK, or equivalent model orchestration frameworks.

-   Deep understanding of prompt engineering, RAG pipelines, semantic embedding, and LLM evaluation.

-   Experience in designing structured outputs from unstructured text via Pydantic or schema-based validation.

-   Strong grounding in HR-tech or NLP problem domains (resume parsing, text classification, dialogue systems, bias analysis).

-   Ability to balance accuracy, interpretability, and cost-efficiency in LLM system design.

-   Familiarity with FastAPI, Docker, Kubernetes, and cloud-based model deployment.

-   Excellent communication and documentation skills, especially for explaining LLM behavior and decision rationales.