ShahebaazSajidKazi
I architect and ship production software using AI as a force multiplier — not a crutch. Faster delivery. Real systems. Still running.
Drag-and-drop, for real
43/43, ticking live
Schema, wiring itself
The actual prompt
Still running
Time vs. team size
05 Systems Deployed to Production
Clinical Data Anomaly Detector
LiveEnsemble Machine Learning Pipeline for Clinical Trial Datasets
Decoupled asynchronous analysis system running an Ensemble ML pipeline to identify univariate/multivariate outliers and physiological trial inconsistencies.
- →Developed an ML Ensemble Pipeline executing 6 detectors: Isolation Forest, Local Outlier Factor, Autoencoder neural networks, Z-Score/IQR statistical estimators, physiological rulesets, and data integrity profiles.
- →Created a Semantic Term Alignment Engine with RapidFuzz fuzzy matching, allowing researchers to upload trial CSVs with arbitrary headers and dynamically map them to standard parameters.
- →Designed an Apple-inspired minimalist cream & black dashboard with bento-grid layouts showing data quality rings, timeseries anomaly spreads, and AG-Grid logs.
Synapse PMS
LiveProject Management Built to Clinical Research Compliance Standards
Replaced fragmented task tracking across WhatsApp, email, and spreadsheets — now the single coordination system used daily by the entire Nucleovir team for all clinical project management.
- →Architected a multi-table PostgreSQL schema using raw SQL + pg.Pool — covering projects, tasks, threaded comments, document versioning, workflow approvals, and team capacity — with RBAC enforcing 4 roles: Admin, Lead, Member, Stakeholder.
- →Built automated notification infrastructure via Resend API: overdue task alerts, daily digest emails, and in-app toast notifications (Sonner) for assignments, due dates, and approvals.
- →Delivered full Admin Control Center: batch task assignment, workload balancing, audit log viewer, user lifecycle management, and clinical domain email enforcement for institutional compliance.
JobTrackr
LiveAI-Powered Pipeline Manager with Predictive Interview Prep & Match Telemetry
Designed and built an AI-assisted job pipeline manager, resolving NextAuth Edge/middleware conflicts and delivering zero-config fallback simulation modes.
- →Implemented Edge-Compatible NextAuth.js v5 (Auth.js) by separating configuration contexts to avoid Edge middleware dependency crashes.
- →Built resilient environment-aware fallbacks for AWS S3, Resend, and LLMs (Claude/GPT) to allow zero-config offline local developer testing.
MealFlow
LiveInternal Operations Portal with Full Role Hierarchy
Replaced a manual WhatsApp-based meal ordering process — handles daily orders for the full Nucleovir team, eliminated end-of-month salary deduction disputes, and gave accounts a real audit trail.
- →43 out of 43 Pytest tests passing — covering authentication, CRUD, role permissions, and all edge cases across 4 user types: Admin, CA/Accounts, Employee, and Intern. Zero broken builds.
- →JWT code-only authentication (no passwords, 7-day token expiry) with Pydantic input validation enforcing bounded text, quantities, dates, and money values across all endpoints.
NucleoPortal
In ProgressReplacing 6 Operational Spreadsheets with One System
Consolidating Procurement, Payroll, HR, Leave, Performance, and Inventory — previously managed across Excel files, email threads, and WhatsApp groups — into one platform on a self-hosted Ubuntu VPS.
- →Multiple modules already live: automated payroll slip generation with attendance-based deductions and multi-stage escalation workflows — Day 2 CA alert, Day 3 Director escalation via email and in-app notifications.
- →PO lifecycle tracking live (Draft → Sent → In Progress → Received → Paid → Closed); performance dashboards with auto-calculated composite scores actively in use.
What You're Actually Getting
Not a developer who needs handholding. A builder who ships, takes complete ownership, and operates with zero safety net.
Production Systems Running
Deployed at a real clinical research company. Real users. Real stakes. Still running.
Solo. No Senior Dev. No Safety Net.
No playbook. No team. Just requirements, architecture decisions, and execution.
Tests Green. CI Passes. Ships on Friday.
I write backend test suites and set up CI pipelines. Code quality is not optional.
Production Incidents
Every system I shipped is still running. Every test still passes. That's the standard I hold myself to.
This Is How I Work
A real prompt. A real system. Running in production.
“”
- ✓Multi-table PostgreSQL schema, raw SQL + pg.Pool
- ✓4-role RBAC enforcement (Admin, Lead, Member, Stakeholder)
- ✓Clinical audit logging & tracking pipeline
- ✓Active in production coordination at Nucleovir
I Use AI the Way Senior Engineers Use Their Entire Stack.
Every senior engineer uses tools that abstract complexity — frameworks, ORMs, and cloud services. I use AI the same way: to handle the mechanical parts of development so I can spend more time on architecture, decisions, and quality. The systems I've shipped aren't “AI-generated.” They are designed, tested, and owned by me. AI is one tool in that process.
3 production apps. 7 months. 0 senior developers. 1 developer with the right tools.
Faster Delivery
I scaffold and ship features in a fraction of traditional time. Your sprint velocity goes up.
Lower Time-to-Market
One developer with the right AI tools compresses what typically takes a 2-3 person junior team.
Adaptable to Any Stack
Shipped in Next.js, FastAPI, Node.js, and PHP in the same 7 months. I bridge gaps fast.
Skills & Technologies
Not just names on a list. Hover over the tags to interact. Each tech was used to architect and ship something real.
My Story & Timeline
How a machine learning background and AI coordination shipped 5 production portals solo.
My background is AI & ML — which means I learned to think about systems before I learned to write them. When I joined Nucleovir Therapeutics as a founding team member, that problem-first orientation turned out to be exactly the right preparation.
When I joined Nucleovir Therapeutics on Day 1, I was handed a blank slate and told: build us everything. No senior developer. No playbook. Just requirements, a laptop, and 7 months. I figured things out the only way that actually works — by building real things with real stakes.
Every system I built is still running. Every test I wrote still passes. That's the standard I hold myself to — not “it works on my machine” but “it works in production, under real load, for real users.” That's the developer you get when you hire me.
AI Intern → Technical Lead
Joined as an AI intern. Became the sole technical architect for all company software. Architected, coded, and deployed all three company production portals.
Ethical Hacking Intern
Conducted penetration testing, vulnerability assessments, and exploit developments. Developed practical skills using Kali Linux, Nmap, Wireshark, and Metasploit.
Diploma in AI & Machine Learning
Specialized in Machine Learning, Neural Networks, Python for Data Science, Computer Vision, Natural Language Processing, and Data Structures & Algorithms.
Initiate Contact
I respond within 24 hours. Call, text, or email directly.