Mohammad Asif · Full-stack engineer & AI consultant
I build AI agents, automations and the apps around them for founders and teams — then stay on to keep them running and getting better. One senior engineer who covers the whole stack: mobile, backend, cloud and AI, with 19+ years of shipping behind it.
Real applications for real businesses. If yours isn't listed, ask — these are starting points, not a menu.
Phone agents that sound natural, understand callers in Urdu, English or both, and let people interrupt mid-sentence like they would with a person.
Chat assistants that answer from your own documents and data, and say "I don't know" instead of making things up.
Agents that take over repetitive, multi-step work across your tools, with a human in the loop where it matters.
Native apps built and shipped end to end, from the first screen to the App Store.
Web products with the boring-but-essential parts done right, and AI built in from day one.
For apps that are unstable, abandoned or stuck. I've taken over broken products before and got them shipping again.
The same loop on my whiteboard, applied to your project.
A call to learn the workflow or product, find where AI actually helps, and agree on a clear scope.
Working pieces early and often, so you see progress and steer it — not a big reveal at the end.
Deployed on your infrastructure, documented, and handed over so your team understands it.
Monitoring, tuning and new capabilities as your needs grow. AI systems get better with care.
Pick what fits — we can change it as the work changes.
One agent, one automation, one app — scoped up front, delivered and handed over.
A senior engineer on call every month: new agents and automations, improvements, fixes and advice as you need them.
Reviews, roadmaps and hands-on guidance for your own team, by the session or by the month.
Each one in the same four steps: the problem, what was built, what shipped, and what comes next. Click a row.
Turn a messy universe of websites, SEC filings, news and LinkedIn into a clean, searchable dataset of family offices — automatically.
Independent agents for discovery, classification, enrichment, validation and indexing. No central orchestrator: each agent claims work from the database and passes it on.
RAG search that drops weak matches instead of guessing, plus a research agent that chains tool calls to answer multi-step questions.
Every confidence score traces back to a specific check, and unreachable sites never drop a record — later agents fill the gaps.
Let customers order by phone, in Urdu, by talking to an agent instead of a menu tree.
Asterisk ARI for calls, Deepgram streaming speech-to-text, LangGraph + Claude for the conversation, and RAG for product search.
Google Cloud TTS replies with barge-in, so callers can interrupt like they would with a person.
Open source, so anyone can build on it.
Bring AI into how a VPN company builds products and supports customers, not just into a demo.
Introduced AI-driven development practices, deployed an AI-based chat support system, and built custom agents for complex workflows.
AI now part of daily engineering and customer support.
Driving agentic AI adoption across the rest of the engineering organisation.
The iOS app was in a broken state, there was no macOS app, and the backends needed a modern home.
Rebuilt the iOS app, built the macOS app from scratch, and led a redesign across iOS, macOS and Android.
Fivefold growth in both user engagement and in-app purchases after release.
Moved backends to AWS and brought in Ansible so infrastructure could be rebuilt the same way every time.
Add new VPN protocols to the iOS and macOS apps. Very little help was available online.
Learned Network Extensions, packet tunnel providers and NEVPNManager in depth, then integrated both protocols alone.
OpenVPN and WireGuard live in the FastestVPN iOS and macOS apps.
Grew into owning protocol engineering for the whole product.
Servers would quietly stop responding on some protocols, or connect without letting users browse.
A vanilla JavaScript bot that checks every server across all protocols, confirms browsing works, and runs speed tests.
Instant WhatsApp alerts on failures, and speed-test history in a database to spot weak servers.
Next step: move it onto AWS.
Nineteen-plus years of shipping, from PHP and the iPhone SDK 3.1 to VPN protocols and agentic AI.
Native Apple platforms at the core, cloud underneath, agents on top.
Agentic AI projects on GitHub.
Real-time Urdu voice ordering agent. Asterisk ARI, Deepgram streaming STT, LangGraph + Claude, RAG product search and Google Cloud TTS with barge-in.
View repository ↗ engasix / personal-ai-employee-hackathonA local-first autonomous agent that manages personal and business work around the clock, using Claude Code, an Obsidian knowledge base and MCP servers. Python.
View repository ↗ engasix / foiAutonomous agents that discover, enrich, validate and index family office data with no human in the loop, then serve it through grounded RAG search and a tool-calling research agent. Python, FastAPI, ChromaDB.
View repository ↗Tell me in a few lines what's slow, manual or broken in your business. I'll give you an honest view on whether AI is the right fix, and what it would take to build.
Or book a call directly. I reply within 24 hours.
Free. We go through the problem and whether AI is the right answer.
Scope, timeline and cost in writing, so you know exactly what you're getting.