D
AI Engineer
DataKite
3 hours ago
Full-time
On-site
Kentucky, United States
Join our AI team to deploy and fine-tune open-source LLMs on bank premises, integrate voice models, and build the governed AI execution flows that power MyMoney and Rawi in regulated banking environments.
Turn state-of-the-art LLM and voice research into reliable, on-prem AI products for the financial sector.
What you will work on
Deploying open-source LLMs on-prem with vLLM, TGI, or Ollama Fine-tuning with LoRA/QLoRA and instruction tuning on banking datasets Integrating Speech-to-Text and Text-to-Speech for Arabic and English Building governed AI execution flows with LangGraph, LangChain, or in-house orchestration Implementing tools, function calling, and RAG against bank data and APIs Designing guardrails and safety checks for skills in regulated settings Model evaluation, benchmarking, and quality regression testing Strong candidates have
Bachelor’s in CS, Software Engineering, AI/ML, Data Science, or related field 2–4 years of experience; strong Python with PyTorch, Hugging Face, or TensorFlow Foundations in ML and deep learning (training vs. inference, embeddings, metrics) Awareness of LLM applications: prompting, RAG, fine-tuning Familiarity with REST APIs and backend integration concepts Bonus: open-source LLMs (Llama, Mistral, Qwen), orchestration frameworks, Whisper/TTS, vector databases, Docker, Kubernetes/OpenShift
#J-18808-Ljbffr
Deploying open-source LLMs on-prem with vLLM, TGI, or Ollama Fine-tuning with LoRA/QLoRA and instruction tuning on banking datasets Integrating Speech-to-Text and Text-to-Speech for Arabic and English Building governed AI execution flows with LangGraph, LangChain, or in-house orchestration Implementing tools, function calling, and RAG against bank data and APIs Designing guardrails and safety checks for skills in regulated settings Model evaluation, benchmarking, and quality regression testing Strong candidates have
Bachelor’s in CS, Software Engineering, AI/ML, Data Science, or related field 2–4 years of experience; strong Python with PyTorch, Hugging Face, or TensorFlow Foundations in ML and deep learning (training vs. inference, embeddings, metrics) Awareness of LLM applications: prompting, RAG, fine-tuning Familiarity with REST APIs and backend integration concepts Bonus: open-source LLMs (Llama, Mistral, Qwen), orchestration frameworks, Whisper/TTS, vector databases, Docker, Kubernetes/OpenShift
#J-18808-Ljbffr