Muhammad TaqiHaider
1.5 years shipping production backend systems that don't just work — they perform.
// about.me
Built on fundamentals.
Performance-first backend engineer. 1.5 years in production. Still obsessed with the craft.
“I get a weird kind of satisfaction when a slow API becomes fast.”
At QBS Co., I profiled a high-traffic monitoring endpoint, identified the exact bottleneck, introduced Redis caching — response time went from 1.6s → under 600ms. Clean problem, clean solution.
Over 1.5 years across NestJS and .NET Core, I've shipped distributed systems, async pipelines, geospatial backends, and ML inference microservices — always with measurable outcomes.
Core Stack
Work Highlights
- ▸Cut API response time from 1.6s → <600ms via Redis caching + profiling
- ▸Designed distributed auth: OAuth2, JWT, 2FA, RBAC over gRPC & Docker
- ▸Rebuilt Azure document pipeline sync→async — 70% processing time reduction
- ▸PostGIS geospatial query restructuring + indexing — 25% faster under load
Personal Project
LiveTrackWise — AI Finance Platform
56-endpoint NestJS/PostgreSQL backend · Angular PWA · Flask ML microservice on Hugging Face — transaction categorisation 88% accuracy, anomaly detection AUC-ROC 0.87. Solo. End-to-end.
View Project🎯 Open to Backend & Full Stack roles — remote or Karachi-based. Looking for teams where the quality of architecture actually matters.
// work.experience
Where I've shipped.
Production systems, distributed backends, and measurable outcomes across 1.5 years.
Associate Software Engineer
Feb 2026 – Present- ▸Cut API response latency from 1.6s to under 600ms by introducing Redis caching for third-party responses, eliminating redundant external calls on high-traffic monitoring screens.
- ▸Built SSR dashboards in Next.js/TypeScript surfacing 10 real-time KPIs across employee activity and live monitoring feeds, deployed as part of a production SaaS platform.
- ▸Designed and enforced authentication architecture across 5+ microservices covering JWT issuance, refresh rotation, and session invalidation with zero reported auth breaches in production.
Trainee Software Engineer
Nov 2025 – Jan 2026- ▸Replaced a manual admin workflow by building an RBAC/IAM interface in Next.js managing 160+ permissions across workspace-level feature gates with bulk assignment support.
- ▸Delivered two production backend services: an inventory management API and an SLA workflow orchestration service using gRPC and Docker for distributed inter-service communication.
- ▸Implemented OAuth2 and 2FA across distributed microservices with token/refresh handling and RBAC; integrated WhatsApp Business API for automated live ticket delivery.
Associate Software Engineer
Mar 2025 – Oct 2025- ▸Reduced per-request DB hits from 6–8 queries down to 2–3 by refactoring service-layer logic to consolidate redundant calls across multiple endpoints, preserving full API backward compatibility.
- ▸Eliminated a document generation bottleneck by building an async PDF pipeline with Azure Queue Storage and SignalR, cutting processing time from ~1.6s to 400–500ms per document.
- ▸Improved geospatial query performance by 25% via PostGIS spatial extensions, optimised spatial indexing, and Google Maps API integration on a multi-tenant SaaS platform.
.NET Intern
Dec 2024 – Mar 2025- ▸Built distributed microservices using Docker, gRPC, and CQRS, delivering Workspace Management, Live Monitoring, and Employee Tracking modules with 25% improved inter-service performance.
- ▸Independently developed Product and SLA services; implemented OAuth2 authentication with 2FA and a Neo4j-based recommendation engine with WhatsApp automation using multi-hop graph traversal.
// featured.projects
Things I've built.
Production systems, personal projects, and everything in between.
VersEye
Enterprise AI-Powered Workforce Intelligence Platform
Built frontend and backend integrations for a workforce monitoring platform — implemented SSR dashboards, metadata APIs, and secure role-based access controls with a focus on low-latency UX and reliability.
Key Features
- ProblemMonolithic monitoring caused repeated DB calls and slow, unresponsive dashboards.
- ChallengesIntegrating 10+ distributed services, enforcing RBAC across roles, and delivering sub-second UX for live video feeds.
- ActionsImplemented Redis caching for metadata, refactored API aggregation to eliminate redundant calls, and implemented RBAC/JWT auth with refresh rotation.
- ResultCut API latency and DB hits significantly; reduced onboarding time by ~40% and improved monitoring responsiveness in production.
- LearningsGained experience in caching strategies for streaming metadata, distributed auth design, and systems-level performance debugging.
Tech Stack
Ajeek-KSA
Service Technology Platform
Contributed to a multi-tenant service platform for marketplace operations and SLA enforcement — focused on reliability, observability, and automation of compliance workflows.
Key Features
- ProblemManual service workflows and fragmented SLA tracking caused delays and compliance risks.
- ChallengesModeling complex tenant-specific business rules and ensuring reliable SLA enforcement at scale.
- ActionsBuilt a .NET microservice for product operations and an SLA orchestration service with automated compliance monitoring and async processing.
- ResultStreamlined operations, reduced manual effort, and improved SLA compliance and visibility across tenants.
- LearningsImproved multi-tenant design patterns, async pipeline reliability, and production observability.
Tech Stack
TenXed
Multi-Role Education & Engagement Platform
Contributed to a production-ready .NET 8 platform serving parents, children, and instructors — focused on secure authentication, modular service architecture, notification pipelines, and payment integrations.
Key Features
- ProblemSupport multi-role workflows with secure onboarding, real-time notifications, and payments.
- ChallengesRole-specific UX, reliable async notifications, and third-party integrations (Stripe, Azure Blob, FCM, SendGrid) without tight coupling.
- ActionsApplied Clean Architecture; implemented JWT/RBAC with OTP flows; added queue-triggered Azure Functions for multi-channel notifications; integrated Stripe payments; enabled Serilog and App Insights.
- ResultDelivered a scalable, auditable platform with reliable background processing and secure multi-role access.
- LearningsStrengthened expertise in enterprise .NET patterns, Azure Functions, notification pipelines, and payment integrations.
Tech Stack
TrackWise
AI-Powered Personal Finance Manager
Final Year Project — full-stack AI-powered personal finance platform with a modular NestJS/PostgreSQL backend (56 endpoints) and a self-built Flask ML microservice hosted on Hugging Face Spaces, delivering real-time transaction categorisation, next-month expense forecasting, and anomaly detection.
Key Features
- ProblemManual expense tracking and noisy transaction data hurt categorisation accuracy and delayed anomaly alerts.
- ChallengesDeploy three production ML models with low-latency inference, avoid Hugging Face cold starts, and return results inline.
- ActionsBuilt a Flask ML microservice (SBERT + Logistic Regression, Prophet, anomaly detector) on Hugging Face Spaces; wired inference into NestJS with caching and rate limiting; added event-driven email auth flows.
- ResultDelivered inline categorisation (~88% accuracy) and anomaly detection (AUC-ROC: 0.87); enabled next-month forecasting after 3+ months of history; reduced cold-start impact with caching.
- LearningsGained end-to-end ML production experience across training, deployment, inference design, and resilient API integration.
Tech Stack
CourseFlow
University Course Scheduler
Web-based University Course Scheduling System built with Flask and SQLite. Implements Greedy and Backtracking algorithms to automate timetabling, detect and resolve conflicts, and provide an admin interface for schedule management and algorithm performance comparison.
Key Features
- ProblemManual university scheduling caused frequent conflicts and administrative overhead.
- ChallengesDesigning algorithms that resolve teacher/room conflicts while scaling to realistic course loads.
- ActionsImplemented Greedy and Backtracking timetabling algorithms with an admin dashboard for validation and conflict resolution.
- ResultAutomated conflict-free scheduling with performance comparisons and actionable conflict reports for admins.
- LearningsGained experience with constraint-solving and timetabling tools.
Tech Stack
Stock Management System
Inventory Management Platform
Automated inventory management system that eliminates manual stock tracking. Reduces inventory errors by 80% with real-time updates and predictive threshold-based alerts across multiple stores.
Key Features
- ProblemManual inventory tracking led to stockouts and overstocking across stores.
- ChallengesDelivering real-time stock updates and predictive alerts on constrained infrastructure.
- ActionsBuilt a Flask backend with WebSocket support for instant updates and implemented predictive threshold-based alerting.
- ResultReduced inventory errors by ~80% and lowered stockouts by ~65%.
- LearningsLearned real-time WebSocket patterns and predictive alerting for inventory systems.
Tech Stack
Chatbot Interface
AI-Powered Conversational Bot
Bilingual AI chatbot supporting English and Roman Urdu for customer service automation. Served 500+ users achieving 85% satisfaction with faster first-response handling.
Key Features
- ProblemLanguage barriers in customer support limited automation for English and Urdu users.
- ChallengesBuilding robust bilingual NLP and delivering a mobile-friendly chat UI within time constraints.
- ActionsImplemented NLTK-based pattern matching for English and Roman Urdu and built a responsive Flask frontend for mobile compatibility.
- ResultServed 500+ users with ~85% satisfaction and faster first-response handling.
- LearningsImproved bilingual NLP heuristics and mobile-friendly chat UI design.
Tech Stack
Mini Compiler
Educational C-like Language Compiler in Python
A university compiler design project implementing a full pipeline for a tiny C-like language — from lexical analysis through intermediate code generation — built from scratch in Python to demonstrate core compiler construction concepts.
Key Features
- ProblemNeeded hands-on exposure to compiler internals (lexer → codegen).
- ChallengesImplement clean scoping, expression parsing, and control-flow in one script.
- ActionsBuilt a 4-stage Python compiler — lexer, recursive-descent parser, semantic checker, and stack-style IR emitter (PUSH/STORE/LOAD/PRINT).
- ResultEnd-to-end compiler producing tokens, symbol tables, IR, and clear semantic errors.
- LearningsPractical experience with lexing, parsing, symbol tables, and IR generation.
Tech Stack
LUDO Board Game
Classic Board Game with DSAA
Classic LUDO board game with a built-in AI opponent, implemented using circular linked lists for O(1) piece traversal and capture-prioritization heuristics for competitive gameplay.
Key Features
- ProblemImplementing complex circular board movement and efficient state tracking for LUDO.
- ChallengesMaintaining O(1) traversal and handling forks while keeping AI competitive.
- ActionsUsed singly circular linked lists for board state and implemented an AI with capture-prioritization heuristics.
- ResultAchieved O(1) piece traversal and a responsive AI opponent experience.
- LearningsApplied circular linked lists and AI heuristics for efficient game state management.
Tech Stack
Flappy Bird Console Game
OOP-Based Console Game
Console-based Flappy Bird clone built with OOP design principles, a pixel-perfect hitbox system, and persistent high score tracking — delivering smooth 60 FPS gameplay.
Key Features
- ProblemDelivering smooth console-game performance with accurate collision detection.
- ChallengesAchieving pixel-precise collision in a console environment and persisting high scores reliably.
- ActionsEmployed OOP design with a pixel-perfect hitbox system and file I/O for persistent high scores.
- ResultSmooth 60 FPS gameplay with reliable high score persistence.
- LearningsBuilt a reliable game loop, collision system, and persistent high-score storage.
Tech Stack
NLP Sentiment Analysis
Amazon Product Review Analysis
ML-powered sentiment analysis pipeline for Amazon product reviews, achieving 91% accuracy. Applied SMOTE for class balancing and wrapped models in a Streamlit app for instant business insights.
Key Features
- ProblemManual product review analysis delayed insights and missed negative feedback trends.
- ChallengesSevere class imbalance and the need for an easy deployment for non-technical users.
- ActionsApplied SMOTE to rebalance data, trained classifiers, and wrapped models in a Streamlit app for instant predictions.
- ResultAchieved ~91% accuracy and improved minority-class F1 by ~34%, enabling fast business insights.
- LearningsApplied SMOTE and packaged models for quick business insights with Streamlit.
Tech Stack
DSA Practice Repository
LeetCode Problem Solutions
Structured LeetCode practice repository with 200+ problems across 25+ algorithmic patterns in C++ and Python, designed to accelerate interview readiness and reinforce pattern recognition.
Key Features
- ProblemUnstructured interview practice and inconsistent problem coverage.
- ChallengesMaintaining code quality and covering a broad set of algorithmic patterns consistently.
- ActionsCurated and implemented 200+ problems across 25+ topics with clean, versioned solutions in C++ and Python.
- ResultA systematic practice repository that accelerates interview readiness and pattern recognition.
- LearningsImproved problem categorization and reproducible solution practices for interview prep.
Tech Stack
// tech.stack
Tools I trust.
The stack I reach for when shipping production systems that have to perform.
Languages
Frameworks
Databases & Caching
Architecture
Cloud & DevOps
Security & Testing
Integrations & Observability
// certifications
Proof of learning.
Professional certifications, workshops, and competitive programming achievements.
LeetCode Streak Achievements
Professional Certifications
Software Design Patterns: Best Practices for Software Developers
Workshops & Training
// achievements
Competed & won.
Competitive programming and continuous problem-solving practice.
Technical Competitions
KODERZ COMBAT
Advanced to Round 2
Competed against developers solving real-time algorithmic challenges under competitive conditions.
Game of Codes V
Participated & Competed
Solved complex programming problems in Python and C++ during the competitive event.
Competitive Coding Platforms
// testimonials
What they say.
Colleagues and mentors on what it's like working with me.
I had the pleasure of mentoring Taqi as one of my first three mentees on project Ajeek-KSA, and I was genuinely impressed by their clever logic in building business process flows and their passion for learning system design. Their independent work consistently demonstrated fast delivery without compromising quality, showing both skill and dedication. I highly recommend Taqi for any team looking for a talented .NET developer with strong analytical thinking, rapid learning ability, and a commitment to delivering high-quality work.
Usaid Bin Rehan
Associate Software Engineer
QBSCo
Syed Tanees Ahmed
Associate Software Engineer
I had the pleasure of working closely with Muhammad Taqi Haider at QBS, where we collaborated on multiple projects together. Taqi consistently demonstrated strong backend engineering skills, with a clear understanding of…
Muhammad Nabeel Hussain
Software Engineer
I mentored Taqi on multiple projects where he worked on backend development using .NET. He quickly understood requirements, built clear and efficient application logic, and translated business needs into practical techni…
Basit Tariq
Software Engineer
I had the opportunity to mentor Taqi while he worked as an Associate Software Engineer on our service-based projects, Real Estate and 10XeD. He consistently delivered his assigned tasks on time and demonstrated a strong …
Abbas Dossa
Test Automation Engineer
I worked with him on the Ajeek project, where he consistently identified and resolved issues at the earliest stages, helping the team stay on track. Despite it being his first opportunity, he demonstrated a high level of…
// want to see more?
// get.in.touch
Let's talk.
Have a project in mind or want to discuss opportunities? I'm open.
Contact Information
Open to Backend & Full Stack roles — remote or Karachi-based.