About BLUE :

BLUE is our AI-powered CCTV Video Data Lake + Compression + Analytics platform. It transforms raw CCTV feeds into actionable dense captions and KPIs for industries like Retail, Petro Retail, Manufacturing, and QSR.

Role Overview :

We are seeking a Prompt Engineer who can design and refine Vision-Language Model (VLM) prompts to generate detailed captions from CCTV feeds and translate them into business-relevant KPIs.

What You'll Do :

- Write and optimize prompts for dense captioning of CCTV video feeds.

- Translate captions into domain-specific KPIs (e.g., customer queue detection, safety compliance, fuel pump activity).

- Build reusable prompt libraries adaptable across industries.

- Collaborate with AI/ML engineers and product managers to align outputs with business needs.

- Validate model outputs and refine prompts to improve accuracy and reduce hallucinations.

What We're Looking For :

- Experience with LLMs/VLMs and prompt engineering.

- Familiarity with computer vision concepts and real-world video analytics.

- Knowledge of KPIs in industries like Retail, QSR, Petro Retail, and Manufacturing.

- Comfort with Python and modern AI frameworks (Hugging Face, OpenAI, PyTorch).

- Strong analytical and communication skills.