Heena Chandak
Finance Expert
PA, United States
Skills
Data Science
About
Heena Chandak's skills align with IT R&D Professionals (Information and Communication Technology). Heena Chandak appears to be a low-to-mid level candidate, with 4 years of experience.
View more
Work Experience
Graduate Teaching Assistant (under Prof. Bhiksha Raj Ramakrishnan)
Carnegie Mellon University
November 2023 - Present
- Designing homework assignments and rubrics for CMU's flagship Deep Learning course (11-785) Covering MLP, CNN, RNN, Attention, Transformers, GNNs, with practical applications using PyTorch Reproduced milestone papers including ConvNeXt, ResNet, Attention, and set standards for students
Carnegie Mellon University
September 2023 - November 2023
- Implemented nanoGPT language model from scratch and pretrained on 9B token dataset. Finetuned for summarization and question answering on CNN DailyMail and SQuAD datasets.
Machine Learning Intern
VeyTel LLC
May 2023 - August 2023
- Designed & implemented an innovative synthetic data generation pipeline using SAM & computer vision techniques, generating over 10,000 realistic posterior trunk images with skin lesions to address data scarcity Developed and deployed lesion detection & segmentation pipeline using Detectron2, enabling precise analysis of skin lesions to advance skin cancer research through collaboration with Hillman Cancer Center Pre-trained DenseNet121 on 224k CheXpert chest x-rays for healthy vs unhealthy classification. This transfer learning optimized RALE score calculation in subsequent real x-ray models, leveraging pretrained weights to boost accuracy and accelerate convergence
Graduate Research Assistant (under Prof. Matthew Travers and Howie Choset)
Carnegie Mellon University
February 2023 - May 2023
- Improved an automated screw classification system achieving 95% accuracy to reduce manual effort in e-waste recycling by 50%
Business Operations Associate
ZS Associates
February 2020 - November 2022
- Provided data-driven insights into key strategic decisions for a top pharmaceutical client through multiple long and short-term healthcare projects, resulting in a 20% increase in sales and a 15% reduction in operational costs Architected SAS and Python flows on AWS (EC2-S3) with multi-layer automation for efficient data synthesis and change management. Achieved an 85% reduction in time spent on data processing and analysis. Reduced the turnaround time of report generation by 80% by implementing automation using Python Developed Tableau Dashboards & formulated new metrics to represent Sales, Payer, and Promotional activities Leveraged a variety of medical industry data sources, including IQVIA Xponent, National Prescriber Audit (NPA), FIA, MMIT, ConnectiveRx, Knipper Promotion Activity and RelayHealth to inform strategic decisions