Backend Engineer

I build backends that hold up under their own numbers — async pipelines, multi-tenant data models, and AI systems in production. BS in Computer Science from FAST NUCES.

2 YRS Professional experience
139★ OSS repo contributed to
3.80 CGPA · Dean's List ×8

Django REST, Celery, RabbitMQ, PostgreSQL

AWS, Docker, Terraform, System Design

02 — Experience

Professional Experience

SiRiiL

Associate Software Engineer · Django REST — Feb 2025 – Jul 2026
- Designed the database schema and system workflow for an AI-driven agentic workflow platform.
- Built the REST APIs that orchestrate it.
- Integrated AI agents so people and automated components work the same tasks.
- Ran asynchronous file processing on Celery and RabbitMQ.
- Integrated third-party systems including QuickBooks and Stripe.
- Deployed on AWS Lightsail.
- Led a team of interns — assigning work, reviewing code, tracking delivery in Jira.
Bridge of Hopes

Backend Developer Intern · Django REST — Nov 2024 – Mar 2025
- Planned and implemented the system's database from scratch.
- Built a full ERD mapping data flow and relationships.
- Wrote Django models with the relationships and constraints to hold that integrity.
Invision Solutions Inc.

Backend Development Intern · Laravel — Jul 2024 – Aug 2024
- Built REST APIs for product retrieval, cart, authentication and password recovery.
- Implemented JWT authentication with role-based access control.
- Managed migrations, controllers, resources and request validation.

Education & achievements

FAST NUCES

BS Computer Science · 2022–2026

CGPA 3.80

Government Delhi Science College

Pre-Engineering · 2020–2022

93%

Hamdard Public School

Matriculation

87%

Dean's List ×8
Top-20 Intermediate Board, Karachi
Winner — FAST Developer's Day Speed Debugging
95% in HSC-I
Ranked 1st, Grade 1 through Matric

03 — Open Source

Open Source

Merged into vidanov/aws-architecture-diagram-skill 139

Add 3D/isometric diagram support (aws3d + Allied Telesis icon libraries)

+834 / −3 Lines, across 10 files
56 min Opened to merged
139★ Upstream repo

I was drawing AWS architecture diagrams by dragging icons around draw.io one at a time. I found a skill that generates them automatically — but when I asked for a 3D/isometric diagram, it could only produce flat icons.

Rather than work around it, I went looking. draw.io ships an old 3D icon library that is not documented and is not reachable from any menu — I found it by reading their source. I verified every icon, wrote the reference tables, built two working examples, and opened my first-ever pull request.

“Wow, cool enhancement!” Alexey Vidanov, maintainer — merged the same day

What actually shipped

Two verified icon-reference documents — including an honest coverage-gap table, since the legacy icon set predates API Gateway, ECS/EKS/Fargate, Step Functions, SNS, Aurora, CloudWatch and IAM
A second bundled isometric library wired in specifically to fill those gaps
SKILL.md updated across all three distributions (Claude, Kiro, standalone)
A general "Visual Quality" section — straight arrows, no crossing edges, consistent spacing — that improves every diagram the skill produces, not just 3D ones
Two working example diagrams, one built directly in draw.io as proof the icon names resolve in the real application

View the PR LinkedIn post Fork

04 — Featured Work

Featured Work

01

SenseiBI

Nov 2025 – Mar 2026 · Final Year Project · Backend Developer

Ask your spreadsheet a question in plain English. Get a dashboard back.

Upload a CSV or Excel file, ask in plain English, get a dashboard — no SQL, no chart builder.
8 purpose-built LLM agents: insights, causal analysis, executive summary, normalization, semantic, structure, suggestion, trend.
Extended NL4DV pipeline turns a natural-language question into a real chart.
Multi-tenant from the ground up — businesses, members, invitations, role-based permissions.
Celery + RabbitMQ keep ingestion, dashboard generation and email off the request path.

Django REST
React
PostgreSQL
Celery
RabbitMQ
Redis
OpenAI API
LangGraph
NL4DV

+7 more

Built with Shaheer Beig (Agentic AI) and Shaheer Mumtaz (Frontend) .My scope: backend architecture, data pipeline and the agent layer.

Read case study Backend repo Frontend repo

02

Prospects Enrichment Tool

Jul – Aug 2026 · Solo

Most email verifiers throw away good leads. This one knows the difference between a dead mailbox and a blocked check.

Turns a raw scraped lead list into a verified, deduplicated, personalization-ready one.
Live SMTP probe tells a dead mailbox from a blocked check — M365 and Spamhaus block the probe itself.
MX, syntax, disposable, role-address, catch-all, SPF/DMARC and Gravatar folded into a 0–100 confidence score.
Scans each clinic’s site and Google reviews for the exact pain being sold against.
Computes a primary_wedge per lead — which pain to open the email with.
Cost-ordered fallback chain: Places → SerpApi → Apify, on a ranked subset only.
Every dropped lead downloadable with the exact reason it was excluded.

Python
Flask
Playwright
SMTP/DNS
Google Places
SerpApi
Apify

Read case study Repo Email Verification Tool

03

ProFit

Oct – Dec 2024 · Team of 3 · Backend Developer

Members, trainers, inventory and finances — with the schema designed first.

Gym admin platform: members, trainers, attendance, inventory and finances.
Owned the backend — normalized schemas designed for data integrity.
JWT auth with role-based access control.
Payment history and salary handling for both members and trainers.
Real-time analytics over revenue, expense and attendance trends.

Laravel
React
Tailwind
MySQL
JWT

+2 more

Built with Shaheer Beig (Frontend) and Shaheer Mumtaz (UI/UX) .

Read case study Backend repo Frontend repo

Also built

The C++ gym system and ProFit are the same problem, two years apart — console app to production Laravel/React product.

05 — System Design

System Design

I designed the core of a stock exchange — order entry, matching engine, market data — anchored on real Pakistan Stock Exchange numbers: 561 symbols, 400 brokers, 620,763 trades/day. Every architecture decision below traces back to a figure from the QPS or storage estimate.