Data Scientist 3

Role: Data Scientist 3

Candidates Required: 1

Focus: Execution, model development, and refining existing pipelines.

Experience: 4–5 Years

Skillset

Modeling:

Expertise in at least 3 modeling areas (Classification, Regression, Clustering, or Recommendation systems & embeddings).

ML/DL Stack:

Production level model development expertise in deep learning & machine learning frameworks: Pytorch / Tensorflow, Huggingface Transformers, Scikit-Learn, XGBoost / LightGBM.

Optimization:

Expertise in model optimization frameworks (e.g., Optuna, Ray) and inference speed-up techniques. Working knowledge of optimization techniques like PSO, Differential evolution.

Experimentation:

Working knowledge of A/B/n testing methodologies, including power analysis, significance testing, and Bayesian approaches.

NLP & GenAI:

Experience using Hugging Face Transformers. Willingness to learn and experiment with LLM fine-tuning (LoRA) and text embeddings.

Engineering & Data:

Proficient in Python and writing complex SQL. Experience with Snowflake / Sagemaker is a plus.

Cloud Architecture:

Experienced with any of model development ecosystems: Databricks / Sagemaker along with core functionalities (Pipelines, Feature Store, MLFlow, Model Registry, Model Monitor) and optimizing data retrieval from Data warehouse systems (Bigquery / athena / cassandra).

Leadership:

Ability to mentor DS2s and define the North Star metrics for complex, multi-stage data products.

Added Values:

1. Reinforcement Learning (RL) or Multi-Armed Bandits for dynamic decision-making.
2. Experience in Causal Inference to understand the "why" behind metric shifts, not just the "what."

Role Expectations

Primary Role: Strategy & Innovation

ML Areas: 3+ Areas (Adv + RL)

Deep Learning: Optimizing / Architecting

Coding: Python / SQL

Testing: A/B/n Testing