Data/Applied Scientist (Search)

Company Overview

(Company overview was not specified.)

Job Summary

The Data/Applied Scientist (Search) role is designed to enhance search capabilities using advanced machine learning techniques. The candidate will contribute to the strategic goals of the organization by developing, deploying, and optimizing machine learning algorithms specifically focused on search functionalities.

Responsibilities

• Develop and implement machine learning models throughout the entire lifecycle including data preparation, training, evaluation, and deployment.
• Conduct vector search and hybrid search techniques along with query preprocessing.
• Generate embeddings using models such as BERT, Sentence Transformers, or custom models and ensure efficient embedding indexing and retrieval.
• Work with GCP services for machine learning and data science, focusing on deploying models using platforms like Vertex AI, Cloud Run, or Cloud Functions.
• Analyze search relevance metrics like precision@k, recall, nDCG, and MRR to evaluate model performance.
• Collaborate on building end-to-end ML pipelines for search and ranking applications and manage CI/CD practices for model versioning.

Qualifications

• Strong programming skills in Python and experience with Jupyter notebooks and packages including polars, pandas, numpy, scikit-learn, and matplotlib.
• Hands-on experience with the machine learning lifecycle.
• Proficient in SQL and BigQuery for analytics and feature generation.
• Familiarity with embedding indexing/retrieval tools and techniques including Elastic, FAISS, ScaNN, and Annoy.
• Experience with Learning to Rank (LTR) techniques and libraries such as XGBoost and LightGBM.
• Understanding of both semantic and lexical search paradigms.
• Exposure to CI/CD pipelines and model versioning practices.

Preferred Skills

• Familiarity with Vertex AI Matching Engine for scalable vector retrieval.
• Experience with TensorFlow Hub, Hugging Face, or other model repositories.
• Knowledge of prompt engineering, context windowing, and embedding optimization for LLM-based systems.
• Comfort with working on BM25 ranking principles via Elasticsearch or OpenSearch and integrating them with vector-based methods.

Experience

• Specific years of experience required were not mentioned; however, broad familiarity with the outlined skills and responsibilities is expected.

Environment

• (Typical work setting and location were not specified.)

Salary

• (Salary information was not provided.)

Growth Opportunities

• (Opportunities for career advancement were not specified.)

Benefits

• (Benefits offered were not specified.)