Data Scientist

Applaudo Studios

About The Position

You are an experienced Data Scientist with strong applied Machine Learning expertise and a track record of building and evaluating models using real-world, messy, large-scale data. You are comfortable working with embeddings, semantic similarity, LLMs, NLP, classification, and both supervised and unsupervised learning. You approach ambiguous problems through structured experimentation, clearly defined hypotheses, baselines, metrics, and error analysis. You are highly autonomous, intellectually honest about experimental results, and able to clearly communicate technical recommendations and trade-offs to engineering and business stakeholders.

Requirements

5+ years of professional Data Science / Machine Learning experience.
Strong applied Machine Learning fundamentals.
Excellent Python and SQL skills.
Hands-on experience with embeddings and semantic similarity.
Practical experience applying LLMs to real-world problems.
Experience with supervised and unsupervised learning.
Strong experience with classification and NLP.
Working knowledge of neural networks and transformer architectures.
Hands-on experience with TensorFlow, PyTorch, PyCaret, or equivalent ML frameworks.
Experience retraining or maintaining classification models in production.
Strong experimental design and model evaluation skills.
Experience defining baselines, metrics, test sets, and error-analysis processes.
Ability to evaluate model quality and demonstrate measurable improvements.
Strong understanding of scalability and ML inference costs.
Strong English communication skills.

Nice To Haves

Entity resolution, record linkage, or deduplication experience.
Ranking and similarity scoring.
Retrieval, clustering, or candidate-generation techniques.
LLM/embedding solutions designed for cost and scale constraints.
Spark, Snowflake, Databricks, or BigQuery.
Experience with company, domain, website, or firmographic data.
Experience working with multilingual datasets.

Responsibilities

Build and evaluate ML approaches for company/entity matching.
Develop embedding and LLM-based matching approaches.
Develop scoring and ranking methodologies to identify true matches and distinguish them from duplicates, lookalikes, and unrelated entities.
Work with messy data, including names, aliases, domains, websites, firmographic attributes, multilingual records, and data hierarchies.
Define benchmark datasets, metrics, baselines, and error-analysis processes.
Design and execute experiments to validate hypotheses.
Compare LLM-assisted approaches against lower-cost alternatives.
Analyze model behavior, edge cases, and trade-offs.
Consider inference economics and scalability from the beginning.
Communicate experimental findings and recommendations to engineering and business stakeholders.
Independently establish experimental pipelines and research approaches.
Clearly document both successful and unsuccessful experiments.