I work on the parts of a system that are expensive to get wrong.
I’m a software engineer in Lahore, at i2c Inc., working on payment platforms. Most of what I do sits behind the request: batch pipelines that must finish inside a window, caches that must not serve a wrong balance, and instant-rail messages that settle irrevocably the moment you accept them.
I came in through .NET, not Java. Four years of freelance delivery across 20+ countries, then an enterprise desktop platform in the UAE where I spent six months making thingsstop being slow. Useful training, because freelance work teaches you that shipping is a skill separate from coding, and a shop-floor application teaches you that a 40-second startup is a business problem wearing a performance costume.
What financial systems taught me that neither of those did is thatcorrectness under concurrency is a different discipline from correctness. A race condition in a jewellery POS is a bug you fix next sprint. The same race in a payment path is money, and on a rail like FedNow there is no batch window in which to notice.
So the questions I find interesting are narrow and repeatable: what happens if this runs twice? What happens if it half-completes? What does the cache I just added cost, and who pays it? Nearly all of my better work has come fromremoving work rather than parallelising it — which is the one thread connecting a 40-second startup to a 40-minute batch.
I use AI coding agents daily, and I’ve also shipped agentic AI into production — integrating Intello and IBM Watson for automated service and real-time event handling. Both sides of that make me think the leverage is real and the gap is specific: judgment about time, ordering and retries. Which is a fair description of what I spend my attention on. I wrote that up here.
Experience
Software Engineer · i2c Inc.
Backend engineering on payment and billing platforms — real-time rails, high-volume batch pipelines, caching architecture and database performance. Promoted from Associate Engineer to Software Engineer.
- Improved Credit Card Payment Distribution from 30–40 minutes per execution to 6–8 minutes, through range-based processing, card-hierarchy grouping, controlled batching and bounded parallelism.
- Implemented parsing and validation for US Federal Reserve FedNow ISO-20022 messages, and documented 3,000+ parameters with valid values and semantic descriptions.
- Worked on Payment Hub, enabling network integrations for financial institutions and FinTechs across Spring Boot microservices.
- Built a pending-payment scheduler and day-end reconciliation services for payment service providers.
- Extended CCPD and bucket distribution with un-adjustment logic for near real-time payment processing.
- Implemented platform-level concerns: rate limiting, AOP cross-cutting, schema validation and centralized configuration.
- Integrated agentic AI (Intello) and IBM Watson for automated customer service and real-time AMI event handling.
- Reviewed and refactored 800+ KLOC, improving structure, maintainability and performance.
- Wrote automated tests and agent-friendly context files for large web-based modules.
Senior .NET Developer · ThinkerSteps Technologies
Enterprise jewellery management platform — performance, architecture and new business modules.
- Cut application startup time from 30–40 seconds to 5–10 seconds by reworking lazy loading and data access.
- Reduced application size 25–30% via reusable UI components and a unified POS module.
- Implemented KYC authentication and an end-to-end manufacturing module.
Level 2 Seller · Fiverr
End-to-end software delivery for clients across 20+ countries.
- 100% completion rate over 3 years 8 months.
- Owned full project lifecycles — scoping, delivery and client communication.
Teaching Assistant · University of Engineering and Technology (UET), Lahore
Programming Fundamentals and Object-Oriented Programming labs.
- Ran hands-on lab sessions and guided problem-solving for undergraduate students.
- Clarified core OOP concepts and helped students implement practical assignments.
University of Engineering and Technology (UET), Lahore — Main Campus
BSc in Computer Science · CGPA 3.106
Engineering stack
- Java 8 / 11 / 21
- Spring Boot
- Spring Cloud
- Spring Batch
- AOP
- REST / HATEOAS
- Microservices
- Concurrency
- ISO-20022
- FedNow
- Network integrations
- Reconciliation
- Idempotency
- Apache Kafka
- Redis
- JCS
- Rate limiting
- Informix
- MS SQL Server
- MySQL
- Oracle
- MongoDB
- SQL optimization
- C#
- .NET Framework / Core
- ASP.NET
- Entity Framework
- MAUI Blazor
- WPF / WinForms
- Agentic AI integration
- IBM Watson
- LLM APIs
- Agent context engineering
- Eval & validation
- Python / Django
- C++
- JavaScript (React / Next)
- NUnit
- Selenium
- Git
- Maven
- Unix CLI
- LaTeX
Recognition
1st Position — Stars of the Month
Release 25.12 (Engineering) · also placed in the Top Ten
Outstanding Performance Award
For work on Payment Hub, Agentic AI (Intello), CCPD and Bucket Distribution
- Guest speaker — engineering careers and freelancing — Quaid-i-Azam University, Islamabad, 2025
Certifications
- SQL — Basic, Intermediate, Advanced
- Introduction to Cloud Computing
- Introduction to Cloud
- SEO · Freelancing
Contact
Happy to talk about backend architecture, payments, or an interesting problem.
Email is the reliable route. If you’re reaching out about a role, a short note about the system and its constraints tells me more than a job title does.