Ahsan Salman Yousaf

Full Stack AI Engineer | .NET 10 & GCP | Building Agentic Systems & High-Accuracy AI Pipelines

Based in: Lahore, PakistanCurrently: Full Stack AI Engineer, EvoSense LabsExperience: 2 years

Full Stack AI Engineer specializing in agentic systems and high-accuracy AI pipelines. Currently at EvoSense Labs, architecting .NET 10 services and GCP infrastructure. Previously at Techanzy, where I automated 15+ workflows and improved data accuracy to 97%. Passionate about bridging accessibility gaps and driving operational efficiency through intelligent automation.

Experience

Full Stack AI Engineer

EvoSense LabsLahore, PakistanworkSep 2026 – Present

Summary

Building secure .NET and Next.js services on Google Cloud with AI image pipelines and CI/CD.

What I did

CI/CD architecture
API development
AI pipeline engineering
Cloud security and infrastructure

Results

Rebuilt the .NET 10 service generator, reducing service setup time from days to minutes.
Rebuilt CI/CD for .NET 10 service generator with secret scanning and Trivy.
Launched production data-modelling API on Cloud Run with 81 automated schema migrations.
Built C# AI photo pipeline using ONNX Runtime for automated background removal, implementing a queueing system to handle GCP memory surges.
Refined edge detection and outline accuracy for background removal to ensure high-fidelity face clarity.
Implemented secure Next.js platform with Microsoft Entra ID and Cloud Armor.
Reduced service generation time from days to minutes.

C#.NET 10Google Cloud RunCloud SQLPostgreSQLNext.jsGitHub ActionsONNX RuntimelibvipsMicrosoft Entra ID

AI Solutions Engineer

Techanzy LimitedLahore, PakistanworkMay 2025 – Present

Summary

Built 3 Power BI dashboards on CRM and operational data to track key business KPIs, improving reported data accuracy from approximately 85% to over 97% and reducing reporting preparation time by 60%. Developed Python data-transformation and validation scripts for API and cloud-application data, cutting average handling turnaround from 4 hours to under 45 minutes and eliminating recurring manual-entry errors. Automated 15+ end-to-end business and data workflows using Make (Integromat), n8n, REST APIs, and webhooks, saving approximately 30 hours of manual processing per week across client engagements. Used AI-assisted development, LLM prompting, and 8 deployed browser-use agents to accelerate prototyping, debugging, data extraction, and repetitive research, saving an estimated 20 hours of analyst time per week, with human validation before production release.