Engineering Portfolio

A scoring system that turns raw campaign data into a single Green/Yellow/Red health rating per client, so account managers can spot at-risk accounts before a client does. Pulls live metrics from TapClicks and delivery data from Asana into one composite score.

• Weighted scoring model — Campaign Performance 40%, Delivery/Execution 25%, Client Communication 20%, Strategic/Account Risk 15% — rolled into a 0–100 composite
• Hybrid benchmarking: universal month-over-month trend scoring for every client, plus optional industry-specific benchmarks for verticals like legal, medical, and automotive
• Built an auto-scoring pipeline — client discovery → metric pull → scoring → database upsert — end-to-end tested against live client data
• Resolved data-integrity issues along the way: pagination bugs, type mismatches from TapClicks' string-formatted numbers, and placeholder scores masking missing data

A fully autonomous weekly newsletter for East Valley Arizona (Gilbert, Chandler, Queen Creek, San Tan Valley) that researches, writes, and publishes itself — no human in the loop once it's scheduled.

• Runs on a cron schedule every Wednesday morning, generating that week's issue from scratch via the Claude API
• Aggregates local events and listings from Eventbrite, Yelp, Ticketmaster, and Harkins/Cinemark APIs before writing
• Renders finished issues to PDF and persists them to a PostgreSQL backend on Supabase

Client reporting used to mean pulling numbers by hand from multiple platforms after every campaign. This app automates that: it pulls from internal databases and the external email marketing platforms REV77 runs campaigns on, then generates, stores, and lets the team browse finished reports from one dashboard.

A personal NFL fantasy football analytics dashboard, built to give myself an edge in drafting and weekly lineup decisions — and to teach myself web development from a standing start of zero, coming from a Python/MATLAB background. Rebuilt from an early Streamlit prototype into a full Dash application.

• Historical stats (2024–2025): QB/RB/WR/TE rankings with red zone share, opportunity share, YPRR, YBC/ATT, YAC/ATT, and boom/bust game-log coloring
• 2026 draft prep suite: overall and positional rankings, strength of schedule, a mock draft board built on FantasyPros PPR ADP, and a toggleable personal draft-notes section
• Team Analytics tab — pass/run rate, red zone conversion, 3rd-down %, sacks allowed — rendered in an AG Grid with heat-map coloring
• Coaching Changes tab tracking all 10 new 2026 head coaches as color-coded impact cards

Designed and modeled the physical competition environment — a makeshift city navigated by an autonomous robot to collect simulated trash units — placing 1st among competing university teams in a timed head-to-head competition.

Assembled collector unit

On the competition course

Head-to-head against a competing team's robot

• Designed and modeled map components in SolidWorks, building the physical environment the robot used to navigate and collect simulated trash units
• Collaborated with a 4-member set-design team to assemble and construct the competition course, including physical testing and validation of the robot

A 3D-printed and hand-crafted small-scale orangutan habitat with an autonomous, interactive attraction, built with a 4-member team.

Finished build — control panel & wired interior

3D-printed gorilla & banana props

Tinkercad model

• Designed the habitat in Tinkercad and fabricated components through 3D printing and hand-crafted construction
• Calculated model ratios and component dimensions to ensure proper fit for mechanical assembly and 3D printing
• Programmed and integrated Arduino-controlled buttons and an automated trapdoor into a fully functional interactive system