Scalable on-chain infrastructure.
Backed by a16z, IDEO, and Coinbase.

• Manage 5+ strategic accounts by selling and becoming a subject matter expert in technical product capabilities. Utilize product suite and Postman APIs/protocol documentation to design solutions and present proof of concepts (PoCs) to clients.
• Prospect 300+ customers via Email/Twitter/Discord, conduct 30+ discovery calls to gather user requirements, and fully onboard 3+ enterprise-grade clients.
• Add 30+ pages to product and developer documentation, leading to 20% reduction in average customer support ticket length. Liaise with product team 2x/week to share feedback from the market and improve product roadmap.
• Lead 2 successful product launches and 4 feature launches including positioning, growth hacking, and partnership coordination.
• Develop end-to-end off-chain analytics infrastructure (Amplitude and Google Analytics) through JavaScript SDK.

• Investigate $1M market for sales enablement feature within SMB SaaS space and develop market research, product positioning, and go-to-market strategy.
• Partner with 2+ software engineers and product leaders to build end-to-end A.I. prescriptive analytics API feature using Python and AWS.
• Plan launch beta rollout for feature, onboard 15 customers, create estimated $70,000 ARR.

• Develop customer upsell algorithms in Python leading to estimated 3x increase in insurance revenue.
• Predict 40k high-propensity clients with 95% accuracy using XGBoost and SVM models.
• Implement SMOTE resampling and Optuna hyperparameter tuning to improve model recall by 53%.
• Design deployment infrastructure for model production using AWS S3 and EC2.

• Discover $130M market for carbon reduction within supply chain and create document outlining opportunity.
• Collaborate with 5+ data scientists and program managers to build carbon pricing product for internal stakeholders.
• Develop Product Requirement Documents for sustainability, operations, and finance teams to promote cross-functional consensus on product vision.
• Design product roadmap for integration with supply chain optimization algorithms.
• Ship product, onboard 4 teams, accelerate enterprise sustainability goals for 2030 by 22%.
• Win 1st place in Nike Hackathon by building customer experience product with natural language processing and machine learning.

Perform economic data analysis using machine learning models in Python and R.

Prepare 2 peer-reviewed research papers for publication in prestigious economics journals:

“Stylized Trends in 21st Century US Onshore Oil and Gas Drilling Geography from Cluster Analysis.”

• Study geographic development of oil and gas industry across US using cluster finite mixture models fit to county-level time series data.
• Import natural gas data in CSV format containing 1 million rows. Write R scripts to clean data using packages like dplyr, reshape2, tidyverse.
• Apply machine learning algorithms such as finite mixture and clustering models using FlexMix package.
• Prepare research paper for publication using ggplot2, Microsoft Word.

“Dynamics in Drug-Related Disorders from 1999-2018 in the U.S.”

• Examine drug abuse in non-drug overdose deaths in US mortality database from 1999 – 2018.
• Collect CDC mortality dataset in RDS format containing 200,000 rows. Clean in R using packages like dplyr, reshape2, tidyverse.
• Estimate from data analysis suggests severity of drug abuse in non-drug overdoses from 1999 to 2018 was 20% higher than current estimates.
• Prepare research paper for publication using ggplot2, Microsoft Word.

Service Fortune 500 clients with machine learning and advanced analytics.

L3Harris Geospatial Artificial Intelligence Analytics

• Develop research framework for merging machine vision tracklets to improve tracker performance.
• Analyze 1200 frames of data to form 500+ data labels from motion imagery geospatial data sets to produce ground-truth data in CSV format. Perform data quality control using Linux Bash scripts.
• Prepare processed and labeled data for training and testing using CNN models.

• Execute critical decision-making in high-pressure situations.
• Collect data on EMT effectiveness by recording situational triages, medication administration, and patient statuses.

• Support cardiovascular health research using wet lab techniques including polymerase chain reaction, gel electrophoresis, cell imaging, etc.
• Acknowledged in cardiovascular health publication in EMBO Molecular Medicine entitled "MicroRNA-574 Regulates FAM210A Expression and Influences Pathological Cardiac Remodeling" by revising technical acumen throughout the paper

• Develop data-intensive application and scalable machine learning product to predict flight delays and rank airline traffic in real time using flight record data and hourly weather report data.
• Collaborate with data engineers and data scientists to build streaming and batch pipelines using ETL, exploratory data analysis, MLOps, and model monitoring.

• Act as Digital Marketing Analyst performing market research, competitive analysis, and stakeholder communication for company securing Series B funding and expanding into international markets.
• Create pitch deck recommending data analytics vendors for business leadership to enhance marketing tech stack and meet business needs for cloud platforms, tag management, data management, data visualization, customer data platform.

• Examine whether geographic sentiments regarding air pollution correlate with air quality levels using machine learning models in Python trained on data from Twitter and Center for Disease Control (CDC).
• Collect 9000 rows of tweets and county air pollution data from CDC. Clean data using Pandas, NumPy, NLTK. Compare Twitter sentiments with CDC data using clustering and classification models such as K-means, decision tree, and random forest.

The drug overdose epidemic is a significant public health issue, and the federal government and various state and local jurisdictions have implemented policies and interventions to reduce drug overdoses. However, drug users who do not overdose but later die from drug-implicated mental and behavioral disorders are not captured in the majority of mortality reporting and research. Little is known about this population, and yet, it may be growing in importance.

In this paper, we take the first step and describe trends in deaths from mental and behavioral disorders due to psychoactive drug use during 1999-2019 and juxtapose them against trends in drug overdose deaths. We then examine how the rates of these two mortality categories compare at the state level. Finally, we describe demographic characteristics of decedents of mental and behavioral disorders due to drug use and compare them to those of decedents from drug overdoses.

The first two decades of the 21st century witnessed revolutionary changes in America’s upstream oil and gas industry, with a meteoric rise in total production and a shift in where this production takes place. Many researchers have used the shift in the geographic location of drilling to study the effects of oil and gas development on local communities. In this paper we seek, to the extent possible, a stylized representation of this recent geographic shift. Specifically, we ask to what extent the timing of drilling across counties in the contiguous United States over 2000-2019 can be explained by places of old drilling versus places of new drilling, or if there are other categories, e.g. places that have experienced both.

To this end, we employ finite mixture models fit to data on county-level drilling shares by year. We find that two clusters reduce the sum of squared prediction errors by about 41%, while third and fourth clusters add only 7 and 4 percentage points of additional explanatory power, respectively. Our main takeaway is that counties can be placed into two categories, old and new drilling, with relatively little room for additional typologies. Moreover, our approach allows us to spatially identify old drilling and new drilling counties, which we map and compare with well-known oil and gas fields.

GPA: 3.82

GPA: 99.23