I build AI systems that end up actually running the business, not just demoing well on a call. Voice agents handling live calls. Automation that qualifies and follows up on leads without dropping them. Full-stack platforms that connect CRMs, calendars, payments, and dashboards into one working system.
Ten years building web and AI applications. Top Rated Plus, 100% Job Success Score.
Some recent work: a voice system handling live call routing for a regional transit authority, a compliance platform for a real estate business built to survive an actual data-deletion audit, and a legal document engine a law firm still runs today, not a prototype they tried once.
What I build:
Voice and Conversational AI
I build voice agents that handle live inbound and outbound calls, not IVR scripts, systems that actually converse, qualify, and hand off cleanly to a human when needed. Built with Deepgram and Whisper for speech to text, ElevenLabs for text to speech, and Twilio/SignalWire for real time call handling and IVR workflows. One recent build: an AI voice system for a home services company that recovers missed calls within seconds instead of losing them to voicemail.
CRM and Lead Management Systems
I build AI driven lead management on top of the CRM you already run: multi source lead intake, AI lead scoring and qualification, automated follow up sequences, pipeline tracking, and conversion reporting in one dashboard. I've built this directly on Zoho and GoHighLevel, not just generic CRM APIs. The integrations that actually matter, custom fields, pipeline stages, automation triggers, are where most builds break, and where I've spent the most time.
AI Agents and Custom Automation
I design AI agents, multi agent workflows, and custom business process automation that handle real operations end to end: lead qualification, follow up, document processing, customer support triage, and internal ops. Built with custom code and API integrations (webhooks, REST, third party platforms), with human in the loop approval where it matters, using OpenAI and Claude. I architect these as clean, maintainable systems, not brittle scripts. A legal document abstraction engine I built using Claude and Azure Document Intelligence is still running in production today.
Full-Stack SaaS and MVPs
I take validated ideas to working, multi tenant SaaS products: user accounts, dashboards, admin panels, role based access, subscriptions, and reporting. Stack: Next.js, React, Node.js, Nest.js, Python, FastAPI, Supabase, PostgreSQL, Stripe, and cloud deployment on AWS and GCP. A landlord compliance SaaS I built for the Baltimore City market is a good example of what this looks like end to end, milestone based delivery, working build at every checkpoint.
Hamza Y. earns an estimated $14k/mo. That's 9.6× the typical freelancer and more than 99.95% of everyone we track.
RAG and LLM Applications
I build retrieval augmented generation systems and document intelligence pipelines: chat with your documents, knowledge bases, structured extraction, and grounded outputs with tight retrieval constraints to keep hallucination genuinely low. Using LangChain, Pinecone, Chroma, vector embeddings, prompt engineering, and contextual memory across GPT-4, Claude, Llama, and others.
Core skills:
Next.js, React, Vue.js, React Native, TypeScript | Node.js, Nest.js, Python, FastAPI, Laravel | OpenAI, Claude, Llama, LangChain, Pinecone, Chroma, RAG | Supabase, Firebase, PostgreSQL, MongoDB, AWS, GCP | Stripe, Twilio, ElevenLabs, Deepgram, webhooks, API integrations
If you're building something like this, message me with what you have in mind. I'll tell you honestly whether it's a good fit.