Curriculum Vitae

Devin Oommen

Chicago, IL · devin.oommen@gmail.com · 630-456-6784 · LinkedIn ↗ · GitHub ↗

Tools & Skills

Languages

Python R SQL JavaScript VBA HTML/CSS Bash

Python

pandas NumPy SciPy scikit-learn Flask FastAPI faster-whisper asyncio aiosqlite pdfplumber BeautifulSoup Selenium Playwright geopy PyInstaller Pydantic RapidFuzz httpx

R

tidyverse sf / ggplot2 Shiny Leaflet stargazer lmtest sandwich car AER (ivreg) microbenchmark parallel

Statistical Methods

OLS / MLR 2SLS / IV MLE Monte Carlo Simulation Bootstrap CI Permutation Tests KDE Bayesian Inference Log-Linear Models Interaction Effects Heteroskedasticity Diagnostics (BP/White) LPM Causal Inference

ML & NLP

Whisper Speech-to-Text (CTranslate2/Int8) Voice Activity Detection (WebRTC VAD) Fuzzy Text Matching (Aho-Corasick + RapidFuzz) Keyword Detection Pipelines

Data & Infrastructure

ETL Pipelines PDF Parsing Web Scraping (Static + Dynamic) Geocoding (Google Maps API) Geospatial Visualization (Choropleth) Parallel Computing Code Profiling Real-Time Audio Streaming (FFmpeg) SQLite Git / GitHub Azure DevOps Jira Tableau

Reproducible Research

Quarto R Markdown LaTeX Jupyter Notebooks APSA Citation Workflow

Dev Workflow

AI-Assisted Development (Claude Code, GitHub Copilot) Test-Driven Development Async Architecture macOS App Packaging

Work Experience

Northern Illinois University, Dept. of Political Science

2024 – 2025
Graduate Research Assistant
  • Conducted independent honors thesis: merged 12 SHRUG datasets (484,630 villages), ran MLR with log-linear specification, interaction terms, and centered covariates; tested heteroskedasticity (BP/White) and multicollinearity (VIF); generated choropleth maps of consumption and electrification. Presented at MPSA 2025.
  • Built multi-iteration Python scraper for presidential election PDFs (history.house.gov); stateful line-by-line parser handling page boundaries and CONTINUED headings; produced 2,000+ row dataset across 6 election cycles, 50 states, 233 party/candidate entries.
  • Analyzed ACLED conflict event data (cross-tabulation of political violence by actor, event type, region) for comparative case study on Indian development.

Revolution New Medicine

Sep 2025 – Dec 2025
Operations Systems Analyst
  • Designed Power Query pipelines and VBA-driven Excel tools for class-booking, patient-management, and daily operating reports; cut ~10–15 hours of weekly manual work and reduced refresh/version-control issues.
  • Used R to build automated reporting infrastructure and dashboards to expand access to operational data and analytics.

Shaw Local — The Daily Chronicle

July 2025 – Present
Freelance News Reporter
  • Produce news content on deadline; pitch enterprise stories aligned to editorial goals.

Northern Star (NIU)

Jan 2025 – May 2025
News Editor & Data Engineer
  • Built end-to-end crime log ETL: Python scraper for 250+ PDFs, 5+ parser iterations (PyPDF2, pdfplumber, regex, positional columns), geocoding via Google Maps API with 30+ hardcoded campus fallbacks, R Shiny interactive map ↗ deployed to shinyapps.io. Pipeline served live newsroom production.
  • Achieved 155% increase in average daily page views over four months; monthly unique users reached 46k.
  • Investigated fire code violations using FOIA requests; documented 15 months of administrative delays preceding a fire that displaced 60 residents.

Northern Star (NIU)

Aug 2023 – Dec 2024
News Reporter
  • Covered every city council meeting over six months; 1,000–3,000 views per article.
  • Reported on university policy and campus events with interviews and data analysis.

Haroon Atcha for COD Board Campaign

Dec 2022 – May 2023
Data Analyst
  • Built voter targeting model in R using Illinois voter roll; identified likely supporters by filtering for voters who participated in 2 of 3 prior municipal elections and voted in Democratic primaries across 50+ precincts.
  • Built field recommendations and data-driven slide deck supporting 40k+ votes.

Evolve Fitness & Training

Oct 2023 – July 2025
Personal Trainer
  • Led 1–3 group classes weekly; coached safe movement and positive member experience.

COD Courier

Aug 2021 – May 2023
News Reporter
  • Reported on Board of Trustees meetings and campus events; maintained student newspaper website.

Unify One Solutions

Aug 2019 – May 2020
Software Consultant
  • Performed manual and regression testing for Surfacide's Android medical device application; authored the FDA De Novo case study that supported regulatory clearance. Tested across 3–4 clients including Shamrock International.
  • Built web scrapers (Python/BeautifulSoup) for data collection; managed test workflows in Azure DevOps and Jira.

Education

Northern Illinois University

May 2025
B.S. Political Science (Honors)
  • Associate University Honors | Mortar Board Honor Society
  • Key Coursework: Computational Methods in Statistics (R/Python: MLE, Monte Carlo, bootstrap, KDE, parallel computing — closed-book coded exams); Econometrics + Lab (R: OLS, 2SLS/IV, LPM, robust SEs, heteroskedasticity diagnostics); Rational Decision Making/Honors (Bayesian inference, CLT, experimental design); Political Analysis; Discrete Mathematics

College of DuPage

2023
A.A. Liberal Arts (Honors), Python Proficiency Certificate (Honors)
  • Key Coursework: Advanced Python Programming; Intro to Python; Programming Logic & Technique; Computers/Information Systems; Statistics for Social/Behavioral Sciences

Presentations & Research

Determinants of Household Electrification — A Cross-Sectional Analysis of Socioeconomic & Gender Inequities in India

Midwest Political Science Association (MPSA) · 2025

Statistical analysis using SHRUG dataset (600k rows) examining links between electrification and household welfare in rural India. Built multiple linear regression models with interaction terms; identified a positive effect of domestic electricity access (~1.6–3.2% increase in estimated monthly consumption per capita).

Do Voters in Community College Board of Trustees Elections Vote Along Party Lines?

College of DuPage Research Symposium · 2023

Analyzed ISBE campaign finance and election data; found correlation between fundraising and voter turnout. Examined demographic and party affiliation data to uncover trends in candidate success and representation.

Data Privacy & Data Regulations as a Function of Hegemonic Power

College of DuPage Library Research Symposium · 2021

Linked data regulatory regimes to global power structures; examined how hegemonic states shape international data governance frameworks.

Technical Projects

DeKalb Scanner Alerts

Python · faster-whisper · FFmpeg · WebRTC VAD · FastAPI · WebSocket · asyncio · aiosqlite · RapidFuzz · Pydantic · PyInstaller
  • Real-time police scanner transcription and alerting system. Ingests live Broadcastify audio via FFmpeg, runs Whisper STT with Int8 quantization, detects critical keywords via Aho-Corasick + fuzzy matching (90% threshold), pushes alerts via email/Slack/WebSocket dashboard.
  • Fully async architecture (14 modules, ~800 LOC). Packaged as standalone macOS app via PyInstaller.

Follow The Money Illinois  Live Site ↗

Python · Flask · SQLite · Playwright · FEC/OpenFEC API Integration
  • Built an end-to-end public campaign-finance data platform ingesting Illinois state bulk filings and FEC federal data into a unified analytics layer.
  • Automated recurring federal sync + analytics refresh with production deployment on Linux and scheduled systemd timer workflows.
  • Implemented reconciliation views and donor intelligence features, including searchable dashboards and donor/network analytics for high-volume campaign-finance data.
  • Ingested 6.4M+ receipts rows, 530k+ D2 totals, and 1.3M+ donor analytics rows with normalization and quality checks for reliable reporting.
  • Improved responsive/mobile UX and platform reliability through pytest coverage and recurring QA checks.

Crime Log ETL Pipeline  northernstar.info ↗

Python · R Shiny · Leaflet · Google Maps API
  • 250-PDF scraper with 5+ parser iterations, geocoding pipeline, interactive Shiny map.
  • Published monthly in AP-style crime reports with embedded maps.

Legislative Bill Stats — Full-Stack Analytics Dashboard  Live App ↗

Python (FastAPI) · SQLite · JavaScript · Chart.js · D3.js
  • Built a full-stack web application (FastAPI + vanilla JS) that aggregates bill sponsorship, cosponsorship, and enactment outcomes for the U.S. Congress (Congress.gov API) and Illinois General Assembly (ILGA FTP XML).
  • Implemented resilient data ingestion with pagination, parallel fetchers, retries/timeouts, background refresh, and layered caching (disk + SQLite) to manage API rate limits while keeping the UI responsive.
  • Designed an indexed SQLite schema (legislators, bills, cosponsors, laws) and computed derived metrics including public/private law breakdown, time-to-enactment, and bipartisan co-sponsorship scoring.
  • Delivered an interactive dashboard with filter/sort/search, party-level aggregate statistics, leaderboards (most sponsored, most enacted, highest success rate, most bipartisan, fastest to enact), CSV export, Chart.js legislative activity timelines, and D3.js co-sponsor network visualizations (force-directed and hierarchical edge bundling).
  • Added pytest unit and integration tests and a cron-friendly refresh script for automated data updates.

Electrification & Welfare in Rural India  Honors Capstone

R · sf · stargazer · Quarto · LaTeX
  • Cross-sectional analysis of 484,630 Indian villages (12 merged datasets). MLR testing domestic, commercial, and agricultural electrification effects on consumption.
  • Found 1.6% welfare increase from domestic electricity. R² = 0.82. 11 academic citations (JPE, World Development, Energy Economics).

Monte Carlo Simulation Study — ANOVA Robustness

  • Designed Monte Carlo simulation (n=700 iterations) in R to evaluate ANOVA robustness under varying effect sizes and error distributions.
  • Analyzed Type I error rates and statistical power across normal and non-normal (exponential) error conditions.

Presidential Election Data Pipeline

Research Assistant
  • Scraped and standardized presidential primary election data from 100+ state-party PDF documents.
  • Built reproducible data pipeline for professor's ongoing research.

COD Board of Trustees Financial Pipeline

Python · Selenium · BeautifulSoup · pdfplumber · pandas
  • ETL pipeline scraping 20+ months of board meeting packets. Dual scraping architecture (Selenium for dynamic, requests for static), two PDF parsers (budget expenditures + accounts payable), 10+ regex patterns, CSV aggregation.
  • Built for professor's governance research.

Awards & Recognition

2025
John G. & Barbara C. Peters Scholarship for Outstanding Public Service
2025
Jim Killam Mentor Award
2025
Roy G. Campbell "Best Damn Reporter" Award
2024
ICPA News Story of the Year — Honorable Mention
2023
ICCJA Reporter of the Year — 1st Place
2021
COD Library Research Symposium — Honorable Mention
2021
Academic Excellence — COD Political Science

Extracurriculars & Volunteer

Democratic National Convention — Volunteer (2024) Associated Press — Election Stringer (2022) Economic Student Association (2024)