Aparna Komarla Email & Phone Number

Berkeley, CA, US

Santa Clara, California, United States

• Building a dynamical systems model to simulate customer buying behaviors in the client and data center market, diversity in product segments and demand allocations
• Building a time-series model to predict inflections in semiconductor market revenue as a function of macroeconomic indicators

San Francisco Bay Area

• Building a data science for social good company focused on criminal justice reform
• Partnering with the Three Strikes Project at the Stanford Law (SLS) to identify nonviolent offenders eligible for Prosecutor Initiated Resentencing (PIR) or AB 2942
• Building an interpretable, explainable, rules-based and deterministic model to help legal experts in their decision making (funded by SLS).
• Applying principles of human-computer interaction (HCI) to augment as opposed to replace decision frameworks in high-stakes and high-risk subject areas
• Investigating the scope, reach and impact of resentencing laws in the state and federal prison system across the country

San Francisco, California, United States

Stanford Law School - Three Strikes Project (See Redo.io)

Taught an introductory data analysis course at San Quentin State Prison to build computer and data literacy skills for career development and rehabilitation. Designed course syllabus and materials including lessons and projects

San Francisco Bay Area

• (Solidigm is Intel's NAND memory division that was acquired by SK Hynix in December 2020)
• Developed a demand prediction model (with SKU level granularity) using patterns in market and customer behavior. Built CI/CD pipelines – released a time-series forecasting package to PyPI.
• Designed a model that uses kernel density estimations and time-series forecasting to improve inventory allocations and supply planning. Presented this work at the Institute for Operations Research & Management Science.
• Explored developing a firmware log anomaly detector using a BERT LLM to help reduce engineers’ debug time from hours to seconds.

Bay Area, California

• A data science and journalism effort to improve transparency on COVID-19 in California’s correctional system (www.covidincustody.org)
• Built a comprehensive public dataset on COVID-19 in local jails with a team of 60+ student interns
• Uncovered 60,000+ infections and several deaths that were previously unknown through 200+ public records requests to law enforcement and public health agencies
• Developed research methods, code and PowerBI dashboards to extract and analyze data (see GitHub www.github.com/covidincustody)
• Published data reports and op-eds in the San Francisco Chronicle, Mercury News, CalMatters among other publications
• Shared data for the Centers for Disease Control's (CDC) COVID tracker (via a partnership with the UCLA Law Data Project)

Berkeley, California, United States

• Joined a pre-doctoral computational social science program in the Haas School of Business, UC Berkeley and the Stanford Graduate School of Business
• Developed a Python program to read, extract and process communications in Slack channels using text extraction and hashing algorithms.
• Deployed a pre-trained word2vec model to study the usage of language in workplace communication and extract insights about organizational culture. See GitHub repo: https://github.com/akomarla/slack_msg_processing

Folsom, California, United States

• Led the development of a time-series based modeling approach to solve this problem. Developed a nearest neighbors (k-NN) classifier that uses a drive’s firmware statistics (temperature, reads, writes, age, etc.) to.
• Built a proof-of-concept solution on 100,000+ datacenter customer SSDs

Davis, CA

Designed and taught a 10-week Intro to Data Analysis course at the local adult school in the City of Davis. Program was awarded a grant from the CA Dept. of Education for career-readiness & skills retraining

Folsom, CA

• Developed a model to predict the rate of incoming bugs in the engineering development process as a proxy for product health and market readiness
• Solution was recognized by senior VPs as it enables program leads to shift roadmap milestones in a timely manner
• Metric has been standardized as a department-wide product health measure

Davis, California, United States

Exploratory analysis using R Studio and Tableau on the graduation rates of Pell Grant recipients in the University of California system.

Davis, California, United States

• Created simulations of mitochondrial DNA in Trypanosomes using worm-like chains, freely jointed chains and rigid circles, to quantify a relationship between the distance separating DNA molecules and their likelihood of.
• Results showed that the linking probability of the DNA network when simulated using non-rigid circles deviates significantly from the standard model, particularly at small distances.
• Presented my work at BaMBA (Biology & Mathematics in the Bay Area), a computational biology conference at Stanford University, and multiple undergraduate research conferences at UC Davis

Woodland, California, United States

Taught high school and community college mathematics courses. Designed course material including notes, homework assignments and sample exams.