💻 LinkedIn | 📍 Remote Worker

I don't build dashboards. I build data systems — from finding where data lives in multiple source systems to turning it into something useful.

Real-time NBA scores app for Even Reality G2 (glasses) & R1 (ring) smart devices:

    flowchart LR
        subgraph Schedule
            ESPN[ESPN Schedule API]
        end
    
        subgraph Cloudflare
            Worker[Cloudflare Worker]
            Cache[KV Cache TTL 15s]
        end
    
        subgraph Source
            NBA[NBA API]
        end
    
        subgraph Devices
            G2[Even G2Glasses]
            R1[Even R1Ring]
        end
    
        ESPN -->|Triggers every 15s| Worker
        Worker --> NBA
        Worker --> Cache
        Cache -->|Auto-refresh| Worker
        Worker --> G2
        Worker --> R1

Features:

• Triggers Cloudflare Worker every 15 seconds
• Auto-selects active (live) NBA games, falls back to schedule for upcoming from ESPN
• Green glowing "Live Feed" badge when game in progress on webapp
• Play-by-play timeline with monospace alignment
• Tap to cycle through games
• Supports both G2 glasses and R1 ring functionality

May 2025 – Present | Remote

Analytics resource for HRCC (Hilton Reservations & Customer Care) operations, focusing on data efficiency and guest service outcomes.

What I'm working on:

• Building ETL pipelines that combine Salesforce, telephony, and case data to model how contacts get routed and where service breaks down
• Creating Alteryx workflows + Python scripts for anomaly detection — catching unusual patterns in IVR (interactive voice response) behavior before they become problems
• Developing behavioral clustering models to segment caller types and optimize routing logic
• Mentoring other analysts on writing reproducible, testable SQL

Mar 2022 – May 2025 | Remote

This was a full-stack analytics role — I didn't just build reports, I built the entire data infrastructure from discovery to dashboard.

The Cross-Functional Scorecard (CFSC)

Built from scratch to unify workforce performance data across 5+ disconnected systems:

    flowchart LR
        subgraph Sources
            F[Five9]
            P[Paylocity]
            S[Salesforce]
            Ph[Pioneer]
            T[TECSYS]
        end
    
        subgraph Snowflake
            subgraph Raw
                R1[order_events]
                R2[employee_hours]
                R3[customer_cases]
            end
    
            subgraph Transform
                T1[Unified Events]
                T2[SCD Hierarchy]
                T3[Scored Output]
            end
        end
    
        subgraph Looker
            L1[Scorecard Dashboards]
            L2[Employee Views]
            L3[Client Reports]
        end
    
        Sources --> Snowflake
        R1 & R2 & R3 --> T1
        T1 --> T2
        T2 --> T3
        T3 --> Looker

Impact:

• Served 200+ employees with self-service analytics
• Eliminated ~15 hours/week of manual Excel reporting
• Single source of truth for workforce performance

The challenge: When an order moves through the system, multiple employees touch it (PV1 verification, PV2 verification, fill, pack). Traditional reporting couldn't tell you which employee completed which task.

My solution:

    flowchart TB
        subgraph Input
            A[Order URL Token]
            B[Event Timestamp]
            C[Employee ID]
        end
    
        subgraph Process
            D[Window Function]
            D --> E{Is Latest Event?}
            E -->|Yes| F[Link to Order]
            E -->|No| G[Discard]
        end
    
        subgraph Output
            H[Employee Task Attribution]
        end
    
        A & B & C --> D
        F --> H

Used ROW_NUMBER() to identify the last event per task type, then joined back to the order. This let us attribute work to the exact employee who did it — not just whoever was assigned to the order.

Dynamic weighting: Different tasks have different complexity. I built percentile-based scoring (1-5 scale) that normalizes across task types so a fast packer gets compared fairly to a fast verifier.

The challenge: Employees moved between cost centers and supervisors. Traditional joins gave you whoever was currently assigned — not who was responsible on the date the work happened.

My solution:

    classDiagram
        class EMPLOYEE_SCD {
            +int employee_id
            +string email
            +int cost_center_1
            +int cost_center_2
            +int manager_id
            +date effective_date
            +date expiration_date
        }
    
        class HOURLY_DATA {
            +int employee_id
            +date work_date
            +float hours_worked
        }
    
        class ORDER_EVENTS {
            +string order_url_token
            +int employee_id
            +timestamp event_datetime
            +string event_type
        }
    
        ORDER_EVENTS --> EMPLOYEE_SCD
        HOURLY_DATA --> EMPLOYEE_SCD

Built a Type 2 Slowly Changing Dimension (SCD) that tracks every change to cost center and manager assignment with effective dates. Now you can ask "who was this person's supervisor on March 14, 2023?" and get the right answer.

This was a problem no one else at Truepill could solve. It became the foundation for all workforce reporting.

The challenge: Pharmacy fulfillment involves orders → items → prescriptions → claims → inventory → patients → employees. Traditional reports showed one table at a time.

My solution:

    flowchart LR
        O[ORDERS] -->|1:M| OI[ORDER_ITEMS]
        OI -->|M:1| P[PRESCRIPTIONS]
        P -->|M:1| IC[INSURANCE_CLAIMS]
        O -->|1:M| S[SHIPMENTS]
        OI -->|M:1| INV[INVENTORY]
        P -->|M:1| PAT[PATIENTS]
        OI -->|M:1| EMP[EMPLOYEES]

Built a 20+ join Looker Explore that allowed C-Suite see the full journey of a prescription from order to patient delivery with drill downs.

The challenge: Pharmacy system stored user data in nested JSON arrays — doctors had both EMR_CONFIGURATION and PROGRAMS as arrays, each containing multiple values.

My solution: Used Snowflake's LATERAL FLATTEN to explode these arrays into rows:

SELECT 
    D._ID AS USER_ID,
    EMRCONFIGURATION.VALUE:"_id"::STRING AS EMR_PROGRAM_ID
FROM STITCH.SOURCE.DOCTORS D,
LATERAL FLATTEN(INPUT => D.EMRCONFIGURATION:"programs") EMRCONFIGURATION

This let us finally do analytics on which programs each doctor was certified for.

Bachelor's degree, Data Science — University of Maryland Global Campus
Jul 2025 — Expected 2028
GPA: 4.0 Honors: Dean's List (Fall 2025)

LinkedIn Learning:

• SQL for Data Analysis (2026)
• Analyzing and Visualizing Data in Looker (2025)
• Prepare Data for Looker Dashboards and Reports (2025)
• Developing Data Models with LookML (2025)
• Alteryx for Healthcare (2026)
• Alteryx Analytics Tips and Tricks (2026)
• Introduction to Alteryx (2026)
• Data-Driven DEI Decision-Making (2026)
• The Data Analytics of Diversity, Inclusion, and Well-being (2026)
• Auditing Design Systems for Accessibility (2026)

Notion:

• Notion Advanced Badge (2026)
• Notion Workflows Badge (2026)
• Notion Essentials Badge (2026)

Microsoft (GitHub):

• Introduction to Git (2025)
• Introduction to GitHub (2025)
• Introduction to GitHub Actions (2025)
• Manage and configure repositories (2025)

Coursera:

• What is Data Science? — IBM (2023)