Machine Learning Engineer

Chagible; United States; Remote; USD 120,000 - 180,000; Full time

Job summary

Chagible is seeking a role focused on building and operating ML systems with a strong emphasis on adapting large language models for practical use. The candidate will train, evaluate, deploy, and continuously optimize models, and connect experimental improvements to production.

Responsibilities include developing and monitoring models, building data pipelines, designing retrieval workflows, and integrating solutions into internal tools with thorough documentation.

Qualifications

• Strong Python skills and practical experience with PyTorch or TensorFlow.
• Deep understanding of transformer architectures, attention, and LM internals.
• Experience fine-tuning language models (LoRA, PEFT, instruction tuning, or supervised).
• Hands-on ability to build retrieval-augmented generation and semantic search systems.
• Experience evaluating model responses, detecting hallucinations, and scoring quality.
• Deploy models via APIs and work on inference latency, throughput, and cost.

Responsibilities

• Develop, fine-tune, deploy, and monitor ML models.
• Build data pipelines for training/evaluation and run experiments to improve quality and efficiency.
• Create prompt strategies and evaluation methods for LM outputs.
• Design retrieval-augmented generation workflows using embeddings and vector stores.
• Collaborate with annotation teams to improve training data.
• Integrate models into applications and document architectures and decisions.

Skills

Python

PyTorch

TensorFlow

Deep learning

Transformer models

LoRA/PEFT/fine-tuning

Retrieval-augmented generation

Semantic search with embeddings

Model evaluation (hallucination)

API deployment

Tools

Pinecone

Weaviate

FAISS

Embeddings tooling

Vector databases

Job description

Role overview

Build and operate machine learning systems, with a strong focus on adapting large language models for practical applications. The role spans model training and evaluation, retrieval-augmented generation, deployment, and ongoing optimization, connecting experimental improvements to production use.

Responsibilities

• Develop, fine-tune, deploy, and monitor machine learning models.
• Build data pipelines for training and evaluation, and run experiments to improve model quality and efficiency.
• Create prompt strategies and evaluation methods for language model outputs.
• Design retrieval-augmented generation workflows using embeddings, vector stores, and retrieval components.
• Work with annotation and quality teams to improve training data.
• Integrate models into applications and internal tools; document architectures and technical decisions.

Requirements

• Strong Python skills and practical experience with PyTorch or TensorFlow.
• Solid understanding of deep learning, transformer architectures, attention, and language model internals.
• Experience fine-tuning language models, including methods such as LoRA, PEFT, instruction tuning, or supervised fine-tuning.
• Hands‑on ability to build retrieval-augmented generation systems and semantic search using embeddings and vector databases.
• Experience evaluating model responses, including detecting hallucinations and scoring output quality.
• Ability to deploy models through APIs and improve inference latency, throughput, and cost.

Nice to have

• Familiarity with vector database tools such as Pinecone, Weaviate, or FAISS.
• Experience with prompt design and structured output controls.