Python Trabajos a distancia y desde casa en new-jersey ∙ Página 1

47 Trabajos a distancia y desde casa en línea

Data Scientist

None · New Jersey, Estados Unidos De América · Onsite

  • Professional
  • Oficina en New Jersey
Solicitar ahora
Data Scientist -

Key Responsibilities
  • Build and optimize classification, regression, and forecasting models using classical ML and deep learning techniques.
  • Develop and deploy deep learning architectures including LSTMs, transformers, and other sequence-based models for time-series, NLP, and anomaly detection.
  • Design and implement NLP pipelines for text classification, semantic search, summarization, and question answering using transformer-based models (e.g., BERT, T5, GPT).
  • Create RAG (retrieval-augmented generation) pipelines integrating LLMs with vector databases (e.g., FAISS, Pinecone, Weaviate) and document indexing frameworks.
  • Apply and fine-tune LLMs (e.g., OpenAI, Mistral, LLaMA, Cohere) for domain-specific tasks using supervised fine-tuning or LoRA/QLoRA methods.
  • Build and orchestrate multi-agent AI systems using frameworks like LangGraph, CrewAI, or OpenAgents to support tool-using, autonomous agents for decision-making workflows.
  • Collaborate with data engineers, product managers, and stakeholders to translate business needs into production-ready solutions.
  • Mentor and support junior data scientists through code reviews, model design feedback, and collaborative experimentation.
  • Promote best practices in reproducible modeling, responsible AI, and scalable deployment.

Required Skills & Experience
  • 5+ years of experience in data science or applied machine learning, with a strong background in both classical and deep learning methods.
  • Hands-on experience with Python, and libraries/frameworks such as scikit-learn, pandas, PyTorch, TensorFlow, Hugging Face Transformers, and LangChain.
  • Strong understanding of classification metrics, feature engineering, model validation, and hyperparameter tuning.
  • Demonstrated experience with LLMs, including fine-tuning, prompt engineering, and retrieval-augmented generation techniques.