GenAI Engineer

PRIMUS Global Services; Pennington (NJ); Hybrid; USD 120,000 - 190,000; Full time

Job summary

PRIMUS Global Services seeks a Generative AI Engineer to drive production GenAI initiatives in the United States. You will deploy, monitor, and optimize models with a focus on on-premises LLMs, data preprocessing, and scalable AI solutions.

You must excel in Python, LangChain, LangGraph, Streamlit, and vector databases, collaborating with data teams to deliver robust AI applications. The role follows a hybrid schedule with onsite and remote work in the Newark, DE or Pennington, NJ area.

Qualifications

• 4+ years in AI/ML, Python and data engineering with 2+ years GenAI production deployment experience.
• Experience deploying, monitoring and optimizing models in production.
• Familiarity with on-premises LLM deployment, data preprocessing and feature engineering.

Responsibilities

• Deploy and monitor GenAI models in production environments.
• Optimize model performance and latency for real-world use cases.
• Collaborate with data teams to translate business goals into AI solutions.

Skills

AI/ML expertise

Python development

Prompt engineering

Chatbot development

Tools

LangChain

LangGraph

Streamlit

LLMs

Vector databases

Hadoop

ETL

Job description

PRIMUS Job Title: Generative AI Engineer

Work Authorization Requirements: "Are you permitted to work in the United States?"

Work Arrangement: Newark, DE or Pennington, NJ; 3 days onsite and 2 days remote

Required Skills/Experience:

1. 4+ years of combined experience in AI/ML, Python, and data engineering, including at least 2 years of hands-on production-level GenAI experience deploying, monitoring, and optimizing models.
2. Strong Python development skills with LangChain, LangGraph, Streamlit, LLMs, prompt engineering, NLP, chatbot development, and vector databases.
3. Experience with on-premises LLM deployment, data preprocessing, feature engineering, SQL-based data analysis, and production AI/ML solutions; Hadoop and ETL experience preferred.