Deloitte

Jan 2021 - May 2022

Consultant for multiple clients across a broad range of industries.

Projects

Automating COVID-19 Call Center Operations

At the height of the COVID-19 pandemic, I supported the New York City Department of Health's call center IT help desk. The team received hundreds of requests each day, many tied to stale user accounts in an Oracle-based system. Most of this work required manual lookups and repeated triage steps, which created long queues and slow resolution times.

I built a set of standalone Python scripts that automated the identification and removal of stale users. The automation corrected issues in the data funnel, standardized how incoming requests were handled, and replaced hours of repetitive work. After deploying these scripts, triage time dropped by about forty percent and ticket resolution time moved from six hours to under thirty minutes. The improvements exposed additional gaps, which led me to define clearer operating procedures and create documentation that allowed agents to self-service the remaining edge cases.

As the procedures matured, the average time to resolution fell again—from thirty minutes to under five. Call and ticket volume followed the same trend. What had started as roughly nine hundred inbound issues per day fell to fewer than forty by the end of the project. The combination of automation and clear documentation shifted the team from reactive support to predictable, low-overhead operations.

Reflection

The project highlighted how small, targeted systems can remove operational friction when paired with consistent processes. It shaped how I approach support workflows today—focus upstream on data quality, automate high-frequency tasks, and use documentation to convert recurring issues into self-service steps.

Multilingual Campaign Page for California's New Benefits Program

This project started as a green-field effort to introduce Californians to an upcoming statewide benefits initiative. The state needed a public campaign page that could communicate the program's purpose, timeline, and next steps to residents across several languages, all while meeting accessibility and legal requirements.

I worked as one of two developers while balancing other ongoing projects, and I took full ownership of the UI. I built the layout, page structure, and visual components using Tailwind, which allowed for fast iteration as design, content, and legal teams refined messaging. Since the page existed primarily to inform, the emphasis was on clarity, readability, and predictable behavior across devices.

Internationalization became a central constraint. The page needed support for both RTL and LTR languages, variable content lengths, and frequent translation updates. I structured the UI so that shifting text blocks or replacing sections wouldn't distort the layout, which proved helpful as content moved through translators and legal review cycles.

Once the implementation was complete, ownership transitioned to other teams for analytics and distribution. My focus was ensuring the page remained stable, accessible, and easy to update as messaging evolved.

Reflection

The project highlighted how even a marketing page becomes technically complex when multilingual requirements enter the picture. Building for RTL, planning for translation churn, and keeping the interface resilient taught me valuable lessons about designing for global audiences.

Building a Unified Search Platform for Government Agencies

The Department of Defense and partner agencies needed a way to eliminate duplicated work caused by siloed information systems. Our team was tasked with building a centralized search platform that could surface data scattered across multiple internal tools and present it through a secure, consistent interface.

I served as a full stack developer and eventually took ownership of the entire frontend rewrite. This included architecting a new interface using React, Next.js, and Tailwind to support rapid iteration and fast navigation across datasets. The backend stack consisted of Python, Django, PostgreSQL, and Elasticsearch, with the engineering effort shaped by strict security requirements involving top secret access, encryption, and controlled read and write paths. The rewrite enabled us to produce several rapid prototypes that were later used in contract-winning demos.

Beyond implementation, I worked closely with project management, clients, and government stakeholders to scope features and run live demonstrations. Stakeholder meetings required a careful balance, since each participant often represented a different agenda. The challenge was presenting work in a way that delivered clarity to one group without losing the attention of another, which made concise communication and strong time management essential.

The platform provided agencies with a single search interface that reduced duplicated analysis and simplified access to distributed datasets. While formal metrics weren't captured during my time on the project, the system replaced previously manual workflows and made cross-agency information easier to retrieve.

Reflection

This work reinforced how complex secure systems can be, from OAuth flows to multi tenant models and data access layers. It also highlighted the value of selecting the right tools early. The combination of a modern frontend stack and iterative prototypes allowed the team to demonstrate progress quickly and influence future contract opportunities.

Accolades

• Unified Search Platform: Recognized for driving results and improving the user experience.
• New York City DOH: Recognized for driving results and improving the user experience.