2021 to Present - Curiously - Founder & Lead Developer
Curiously is a browser addon in alpha. Curiously is not a cure for cancer, nor is it ground breaking. It is a personal project and a loose rendition of one of my favorite web technologies of the early 2010's: StumbleUpon. I want people to rediscover an internet not governed by tech monopolies, but by the musings and ambitions of curious individuals and laudable organizations.
I am developing Curiously using JavaScript, HTML, CSS, Postgres, and Supabase.
2021 to Present - Boulder County Public Health - Data Scientist
- Manage the development of a case management app for Boulder County Public Health’s regional epidemiologists. The app enables epidemiologists to search and assign cases, log outbreaks, and quality assure COVID-19 patient data more effectively and efficiently.
- Develop and deploy a robust and resilient ELT pipeline using Python, SQL, PowerShell scripts, and Task Scheduler that extract structured and semi-structured COVID-19 data from CommCare and CEDRS, load it into BOCO’s SQL Server, and transform it via stored procedures in SSMS into a usable format for case management app.
- Conduct in-depth interviews with BOCO PH’s epidemiologists and document their needs. Direct BOCO’s App Development team in building the app via mock-ups, technical specs, ERDs, and DFDs.
- Aggregate, cross-reference, and examine data in SSMS, prepare data for Tableua. Create Tableua reports for epidemiologists describing data completeness, case completion, and contact rates.
- Secure sensitive PII data in accordance to HIPAA. Authorize database access for PII authorized users and obfuscate data and create views as needed.
2021 to Present - Numer.ai - Tournament Participant
I submit my predictions from two models every week; a CatBoost and an XGBoost model.

2020 - The Deep Learning Cookbook, Part 1: A Geometric Understanding of Data
Jan 2020 to July 2020 - Machine Learning Engineer – Moove.ai
- Productized a road safety web app that scored the safety of every road within a given bounding box. Built using ML methodologies, Google Cloud Platform, GIS technologies, and multiple structured data feeds. Established robust data and ML pipelines that ingested and transformed billions of rows of data. Product led to a yearly subscription deal with a large tech firm.
- Built an ETL pipeline that ingested, normalized, and scored 1B plus rows of structured data using Cloud Functions and Airflow DAGs. Scheduled DAG runs every 15 minutes that loaded raw and normalized data into BigQuery, then fed normalized data into ClickHouse to rapidly display data in the UI.
- Developed an automated ML pipeline in AI Notebooks that auto-allocated GPU resources in JupyterLab instance, then loaded data, and instantiated, trained, and evaluated the accuracy of road scores.
- Formulated econometric mathematical scoring models using exponential functions, and weighted arithmetic means. Deployed econometric scoring models as UDFs written in JS and SQL in BigQuery.
- Led the design of the data model. Constructed Moove.ai’s data infrastructure with model as the blueprint.
2019 - Dr. Radhika Nath's DPS School Board Campaign

2017 - Dialogues Writer

Dialogues Writer is an intelligent word processor with novel features such as an analogous thesaurus and a story arc mapper. It's features focus on tone, style, and story and character development. If a writer wants to write in a gothic style, then it will suggest words and ideas that guide the writer toward that style.
It is designed and coded using Keras, PyTorch, spaCy, gensim, Flask, Graphene, and Firebase. The story arc mapper graphs the culmulative sentiment of a story, given that each paragraph is assinged a numeric value.
The analogous thesaurus feature is built on a word2vec architecture. It feels similar to a regular thesaurus, yet, instead of synonyms, each word calculated represents an analogous mapping of a word. For example, using the gothic style, if a writer wants an analogous word to church, Dialogues Writer would suggest cathederal.
2017 - 2020 - Dialogues.ai - Analytics Manager
- Managed and worked alongside a team of three machine learning engineers and two full stack developers to productize UI/UX, backend, and API of digital writing assistant SaaS product.
- Worked closely with Product Manager to define product features, managed project using Asana.
- Gathered requirements from key stakeholders, performed in-depth market research, and created and automated detailed KPI reports for each stakeholder.
- Implemented the BERT architecture and trained it using web-scraped and highly preprocessed text dataset. Warehoused trained model and data on GCP.
