aruparnamaity

Senior Data Science Engineer

 Senior Data Science Engineer with expertise in MLOps and end-to-end machine learning solutions. Over my career at Micron Technology and ZS Associates, I have worked extensively on implementing scalable ML models, developing anomaly detection pipelines, and optimizing inventory systems to deliver tangible business value. My core strengths lie in leveraging deep learning frameworks like TensorFlow, PyTorch, and Keras, especially for transformer models and LLMs in production. I’m actively contributing to Large Language Models (LLMs) for material demand forecasting and finetuning models using Supervised Fine Tuning (SFT) and Reinforcement Learning from Human Feedback (RLHF). I also bring strong MLOps expertise, having deployed machine learning solutions on Google Cloud Platform (GCP) with Kubernetes, Airflow, and Docker for automation, monitoring, and model versioning. Skilled in building user-friendly UIs with Streamlit, Angular, and RShiny, I ensure that complex data science workflows are easily accessible and maintainable. With a keen interest in Web3 technologies, computer vision, and deploying ML models for edge devices, I’m excited to drive advancements in these emerging fields.


Experience: 5 years

Yearly salary: $54,000

Hourly rate: $0

Nationality: 🇮🇳 India

Residency: 🇮🇳 India


Experience

Data Science Associate
ZS Associates
2020 - 2022
- Spearheaded the ZS iData research project by designing and implementing an integrated pipeline for anomaly detection across diverse pharmaceutical domains. Deployed solutions in Databricks - Crafted an Intelligent Algorithm Recommendation Engine that autonomously suggests algorithms and probability thresholds for anomaly detection; reducing manual intervention by more than 95%, leading to faster, more reliable insights for the data team - Experienced in writing production level codes, code packaging into installable, and code parallelization - Created complex features and feature engineering using PySpark and SQL - Implemented Positive Unlabeled Learning and Genetic Algorithms for detecting misdiagnosed patients for a rare disease
Summer Research Intern
Indian Statistical Institute
2019 - 2019
- Formulated an innovative mechanism of Garment Transfer from one person to another with a different pose, using Thin plate Spline Transformations. Paper in Springer- Multimedia Tools and Applications - Implementations: Pose estimation using Deep Learning, Semantic Segmentation, Non-affine Transformations, GAN; (tested) Image Inpainting and Shape from shading

Skills

angular
c-plus-plus
data-science
docker
git
machine-learning
nlp
python
sql
stats
ai
english