Engineering systems · AI · Automation

I build systems that turn ideas into execution.

Fahad Malik is an engineer working on autonomous AI systems and automation infrastructure — software that discovers opportunities, makes decisions and executes them without waiting on a human in the loop.

6Systems built
FastAPICore runtime
24/7Builder mindset
Fahad Malik at his desk, working on a laptop with a code editor and a data dashboard visible on dual monitors behind him
Fahad Malik
AI Systems · Automation · Software
● SYSTEM NOMINAL
01 / Live environment

Not a static portfolio.

A simulated engineering workspace, styled after the real loop these systems run on: observe, reason, execute, learn.

spider_engine.py
01async def discover(market):
02    opportunities = await scout.scan(market)
03    for item in opportunities:
04        signal = await analyzer.score(item)
05        if signal > 0.75:
06            await executor.run(item)
07            await feedback.learn(item)
08
09    return system.status()

# observe → reason → execute → learn

Runtime telemetry

Agents active03
Opportunity queue27
Workflow nodes124
12:42:08 system initialized
12:42:11 scout connected
12:42:14 opportunity scored
12:42:17 execution queue ready
02 / Selected systems

Projects with a job to do.

Each one is treated as a system: a real problem, an architecture, an execution loop, and an honest status.

AI / Operating system

SPIDER OS

An opportunity-intelligence and revenue-execution OS. It finds market problems, scores opportunities, qualifies leads and prepares actions, then waits for a human approval before anything goes out.

Problem — finding and acting on business opportunities is manual, slow and inconsistent.
Architecture — FastAPI brain with its own database, a shared Central Nervous System layer, and pluggable business engines. Bayesian scoring, approval gate, kill switches.
Status — ● Live (private)
Python · FastAPI · SQLAlchemy · Alembic · PostgreSQL · Cloudflare Tunnel
Ask about the architecture ↗
Social / Growth engine

Social Media Agent

Runs inside SPIDER. Discovers local-business prospects on Instagram, audits their profile, drafts personalised messages, handles replies and learns which approach wins.

Problem — small businesses need steady, quality social outreach but cannot staff it.
Architecture — Prospecting, audit, drafting, inbox and quote modules; a bandit layer picks the best approach; events flow to the CNS through an outbox.
Status — ● Built, in testing
Python · Instagram Graph API · Bandit learning · Event outbox
Ask about the architecture ↗
Lead engine / Website

Free Business Check

A public form that turns a visitor into a CRM lead and generates a free website and presence check for the owner. The same pipeline now powers the contact form on this site.

Problem — most local businesses never learn how weak their web presence is.
Architecture — Next.js site on Cloudflare Workers, server-side hop with a shared secret, SPIDER lead intake, free check report, owner alert by email.
Status — ● Live
Next.js · Cloudflare Workers · FastAPI · Resend
Ask about the architecture ↗
Product / Education

PadhaiMatch

A managed tutor marketplace for Classes 6–12 in Mumbai, Kanpur and New Delhi. Students post a requirement, verified teachers are matched personally, a demo is booked and payment is collected, with pricing hidden between the two sides.

Problem — finding a trusted tutor, or a student, runs through scattered low-trust channels.
Architecture — Phone/OTP login, document and KYC verification, rule-based matching, WhatsApp notices, payment link and webhook, payout ledger, admin console.
Status — ● Live at padhaimatch.in
Python · PostgreSQL · Supabase · Payments
Visit padhaimatch.in ↗
Content / Automation

AI Content Channels

Agents that plan, script, voice and publish educational video for a client's channel, with the client and the owner approving every video. First live channel: Law with Sana, a Hindi law-explainer channel for LLB students, with a cloned voice (with the creator's written consent), cartoon animation and AI-use disclosure on every video.

Problem — consistent content takes a team most solo creators do not have.
Architecture — One codebase with a client id on every row. Script, voice, video and publishing adapters for YouTube, Instagram and Facebook. A learning loop reads channel analytics and sends topic recommendations to SPIDER. Ships with a kill switch and approval on.
Status — ● Live demo: Law with Sana (first video published 18 Oct 2026)
Python · FastAPI · Open-source voice clone (IndicF5) · ffmpeg · YouTube Studio
Watch the demo channel ↗
Content / Automation

AI UGC Video Ads

Short vertical video ads made with AI video models and a Hindi/English voiceover, written to be honest (no made-up customer claims) and labelled as AI-generated. Ads so far: Spider OS (15 s, 27 Oct 2026), a dental-clinic sample and a doctor-token-system ad (30 Oct 2026), with separate voices for patient, receptionist, doctor and narrator.

Problem — brands need a constant stream of fresh creative and most small businesses cannot produce it.
Architecture — ad research from public ad libraries, script and prompt writing, AI video generation (Veo, Seedance), voiceover (Edge-TTS), captions and labels with ffmpeg, human review before publishing.
Status — ● Three ads made; pilot offers for clinics and small brands open
Veo 3.1 · Seedance · OpenArt · Edge-TTS · ffmpeg
Watch on spideros.tech ↗
Healthcare / Product

QueueCare — Doctor Appointment & Token System

A token system for clinics: patients scan the clinic's QR, get a token on their phone and see how many are ahead. WhatsApp alerts say when the doctor has arrived and when the turn is near, the doctor taps "Call next", and a waiting-room screen shows the number. Sample data only in the demo.

Problem — patients stand in line for a number and clinic staff spend the day managing the crowd.
Architecture — clinic QR and patient web app, doctor desk, reception queue, waiting-room TV board and WhatsApp notifications on one booking core. Runs on its own server; the video above is AI-generated and labelled.
Status — ● Working demo; first clinic pilot in discussion
Python · FastAPI · WhatsApp Business · QR · ffmpeg
Ask for a live demo ↗
Infrastructure

Central Nervous System

The shared layer that lets separate systems see each other without sharing databases: events, metrics, directives and a read-only feed from the marketplace.

Problem — separate engines drift apart and nobody sees the whole business.
Architecture — Its own database, allow-listed read-only feeds with bearer tokens, outbox events from each engine, decisions and learnings recorded back.
Status — ● Live (private)
Python · Event outbox · Read-only feeds · Token auth
Ask about the architecture ↗
03 / Architecture

Observe. Reason. Execute. Learn.

The common pattern behind every system above.

DISCOVERSignals & data
ANALYZEContext & reasoning
DECIDEScore & qualify
EXECUTEActions & delivery
LEARNFeedback loop
04 / Stack

What it's built with.

The tools that actually show up across these systems, not a generic skills list.

PythonJavaScriptHTML / CSSFastAPINext.jsCloudflareSupabaseDockerGit / GitHubPostgreSQLAI / LLM APIsREST APIsVercel
05 / Building in public

The lab is always running.

Experiments are where ideas turn into systems.

● ACTIVE EXPERIMENT

AI Website Auditor

Automated scoring of a business's existing website — speed, structure, conversion gaps — to prioritise outreach.

● ITERATING

Agent Workflows

Agent workflows that research, reason, execute and report back, with a human checkpoint where it matters.

06 / Terminal

Talk to the system.

Try help, projects, stack, about or contact.

fahad@portfolio:~
fahad@portfolio:~$
07 / About

I don't build interfaces. I build systems that keep working after I stop watching them.

I'm Fahad Malik, based in India, working at the intersection of AI, automation and product engineering — mostly in Python, FastAPI and whatever LLM API fits the job.

Most of what I build follows the same loop: discover a real opportunity or problem, decide what's worth acting on, execute it with as little manual work as possible, and feed the outcome back in. SPIDER OS, its social and lead engines and PadhaiMatch are all the same pattern applied to different problems.

I'd rather ship a system that runs on its own than a demo that only works when I'm driving it.

08 / Contact

Have a system worth building?

Tell me what you want built. Your message goes straight into my SPIDER OS pipeline and I reply personally.