Muhammad Talha Bin Zia

With a focus on safety and compliance, this project collects all relevant information and assists employees with safety and compliance as a guide, always available to spread awareness. Currently in progress, this project utilizes open source solution for local deployment.

A generative AI-based system that includes multiple chatbots for assisting small public businesses by automating their press releases, earning call scripts, and a knowledge agent. It uses OpenAI API and Gemini API with Document Vault to answer queries accurately and reliably. I was able to lead and architect the project with my team.

VHT is a third-party patent solution deployed on major contact center accounts, such as the United Health Group, T-Mobile, and Verizon.

It was completely integrated with Afiniti's V6 software by developing a C++ back-end microservice from scratch. It supports a wide range of client customization. A multi-tenant, multi-threaded VHT integration service was implemented. It is responsible for communication between VHT services and Afiniti.

The functionality developed was divided into three major areas:
1. Initializing queue mappings.
2. Pulling queue stats. This involves constantly loading stats from VHT stats servers asynchronously. One thread per server is used.
3. Generating and sending notifications. This involves receiving queue-based events from other Afiniti services.

HA was designed and implemented for parallel communication with multiple VHT servers and shifted to backup servers in case of failure using C++ futures. A gSOAP solution was developed using portable components (POCO) HTTP and pugixml to support SOAP-based communication. SSL and new WSDL support were later added.

I performed a stress testing load of 1,500 calls per minute using an in-house simulator.

SonarQube is a third-party paid service that provides static code analysis across project codebase, identifies security hotspots and potential threats, and suggests best practices.

Based on this analysis, I removed all possible issues across the entire SI-Avaya's codebase.

This activity provided me chance to learn and further improve my understanding.

Dromen Inc. is dedicated to empowering individuals with disabilities by providing comprehensive support services that extend beyond traditional financial management. Their offerings include Financial Management Services (FMS), Non-Public Agency (NPA) services, and Home Health Care, all designed to facilitate clients' participation in the Self-Determination Program (SDP). Emphasizing collaborative, relationship-driven partnerships, Dromen refers to their collaborators as "Community Partners" to highlight a commitment to inclusivity and shared goals. Additionally, they offer training programs for Independent Facilitators to enhance support for individuals with developmental disabilities. Dromen's mission is to transform collective dreams into purposeful realities for each client.

UBPS is a critical component of bill payment infrastructure, processing a high volume of transactions for banks. Self-handled utility companies had all transaction logic implemented in C++ IRIS switch, while third-party telecom companies used SOAP-based APIs consumed using gSOAP for C++. C# code was converted to C++ microservices. I worked with a DevOps team to conduct stress tests. Ultimately, Telenor chose TPS UBPS due to its proven reliability, robustness, and track record, over other top-notch solutions.

Telemetry services help monitor service health by listening on multiple stats.

All required stats, including average round trip times, warnings, and error counts, were calculated and delivered to the monitoring service to be available for support teams across multiple accounts.

I improved the performance of the in-house C++ logger library by 2x. Timestamp generation was reduced from once per log to once per time change. Mid-level conversion from log parameters to log objects was avoided, saving precious string conversion and copying cycles.

Improved the C-based RCoS+ platform for IoT devices. I provided the base functionality to the application teams. The platform comprises an event handler, scheduler, threading support, base devices, and drivers.

ByteDance Piolet Project via Toptal:

Selected as a C++ and problem-solving expert to generate complete and correct test cases for complex programming problems.

Authored high-quality training data with the purpose of enabling a large language model (LLM) to achieve high results on coding challenges.

Recognized as the best performer for being the fastest to complete the assigned quota of problems during the month-long pilot project.