AI Data Scientist

About the company

• Position listed on behalf of a partner company that manages all applications and next steps.
• Research-driven, collaborative environment focused on practical innovation and measurable impact.

Responsibilities

• Research and evaluate emerging techniques in NLP, statistical machine learning, deep learning, and generative AI.
• Independently investigate technical challenges, develop hypotheses, conduct experiments, and compare models and architectures.
• Design and implement NLP and ML solutions using transformer-based architectures.
• Develop data pipelines, preprocessing workflows, embeddings, and training datasets.
• Establish data quality practices to maintain model reliability.
• Build generative AI and LLM solutions using prompt engineering, RAG, embeddings, vector search, fine-tuning, AI agents, and structured generation.
• Optimize model selection balancing accuracy, latency, cost, and business requirements.
• Design experiments, benchmarks, evaluation datasets, and statistical methodologies.
• Conduct error analysis and apply evaluation metrics.
• Move AI solutions from research to production with engineering teams.
• Monitor deployed systems for model drift and performance bottlenecks.
• Provide technical leadership and mentorship.

Requirements

• Master's degree in CS, data science, computational linguistics, statistics, mathematics, engineering, or related field (PhD preferred).
• 10+ years in data science, ML, NLP, or AI with hands-on deployment experience.
• Expert-level Python with Hugging Face Transformers, spaCy, NLTK, Sentence Transformers, PyTorch, TensorFlow.
• Deep knowledge of traditional NLP: tokenization, stemming, lemmatization, n-grams, TF-IDF, topic modeling, text classification, NER, information extraction, semantic similarity, information retrieval.
• Understanding of deep learning architectures: CNNs, RNNs, LSTMs, GRUs, encoder-decoder, attention, transformers.
• Hands-on LLM experience: prompt engineering, RAG, vector search, reranking, fine-tuning, LoRA/QLoRA, structured generation, LLM evaluation.
• Strong experimental design, statistical testing, cross-validation, hyperparameter optimization, error analysis.
• Experience with evaluation metrics: precision, recall, F1, ROC-AUC, MAP, MRR, NDCG, groundedness, relevance.
• Strong SQL and experience with Spark, Databricks, or BigQuery.
• Familiarity with GCP, AWS, or Azure and ML infrastructure: MLflow, Kubeflow, Weights & Biases, Docker, Kubernetes, Ray.
• Experience with vector databases: FAISS, Elasticsearch/OpenSearch, Pinecone, Weaviate, Milvus; frameworks like LangChain, LlamaIndex, DSPy.
• Excellent communication and collaboration skills.

Preferred qualifications

• NLP for legal documents, document classification, semantic search, document intelligence, multimodal models, knowledge graphs.
• Expertise in model quantization and distillation.