09/04/2026 updated


100 % available
Full Stack Developer | AI Agents & Automation Specialist
Lahore, Pakistan
Worldwide
Master of Computer Science (MSCS)About me
Full Stack Developer & AI Automation Specialist, 10+ years experience. I build scalable web apps, SaaS platforms, and production-grade AI systems (agents, RAG, voice automation) not demos, real systems that run reliably. I combine solid engineering with AI to solve real business problems.
LangChainJavaScriptPHPArtificial IntelligenceSoftware as a ServiceDjangoPython (Programming Language)LaravelMachine LearningNatural Language ProcessingNode.jsWorkflowsMulti-Agent SystemsGenerative AIBackend
10+ Years Experience | 250+ Successful Projects | Trusted by Fortune 200 Companies & Funded Startups
Over the past 10+ years, I've built scalable web applications, SaaS platforms, and more recently production-grade AI systems, carrying each one from architecture through to deployment.
My work sits at the intersection of two specialties: solid full-stack engineering, and modern AI agent, RAG, and voice automation. Most clients today don't need one or the other they need a real product with AI woven into it from the start, not bolted on afterward. That's the gap I fill.
What I Build
Full Stack Web & SaaS Development
From multi-vendor e-commerce platforms to subscription-based SaaS products and headless CMS builds, I work primarily in Laravel, PHP, React, Vue.js, Node.js, Python, Django and WordPress, choosing the stack based on what the project actually needs, not habit.
AI Agents & Multi-Agent Systems
I design agentic workflows with LangGraph, CrewAI, LangChain, and MCP systems that can reason across multiple steps, use tools, retain memory, and carry out tasks with minimal supervision.
RAG & AI Knowledge Assistants
When a business needs its own data turned into answers, I build retrieval-augmented systems on LangChain, Pinecone, Qdrant, and pgvector semantic search, document Q&A, and citation-grounded responses that stay accurate to the source.
Voice AI Agents
Using LiveKit, Vapi, Retell AI, and Deepgram, I build real-time conversational agents that handle support calls, book appointments, and qualify leads without sounding like a script being read aloud.
AI-Integrated Web Systems
Rather than treating AI as a bolt-on feature, I integrate it directly into CRMs and business systems — lead scoring, automated responses, and dashboards that actually think, not just display data.
Tech Stack
Backend: Laravel · PHP · Node.js · Python · FastAPI · Django · REST APIs · GraphQL
Frontend: React · Vue.js · Angular · TypeScript · Next.js · Tailwind CSS
AI & Agents: Claude API · OpenAI API · LangChain · LangGraph · CrewAI · MCP
RAG & Search: Pinecone · Qdrant · pgvector · Embeddings · Semantic Search
Voice AI: LiveKit · Vapi · Retell AI · Deepgram
Databases & Cloud: MySQL · PostgreSQL · MongoDB · AWS · Docker · CI/CD
Why Work With Me
I don't build in isolated pieces. The web application, the data layer, the AI logic, and the deployment pipeline are designed as one system from day one which is why projects tend to move from idea to production without the usual rework. Whether you're starting a SaaS MVP from scratch, modernizing an existing platform, or adding AI to something that already works, I'll scope it around what you actually need, not what sounds impressive on paper.
Send me your current stack, your goal, and your data sources — I'll map out the simplest, most reliable way to get there.
Experience
Professional experience5 - 10 years
Freelancer since2016
Languages
EnglishNative speaker
Project history
Built the AI engine and platform behind Mipria, a French wellness brand offering AI-personalized herbal remedies for humans and pets, blending traditional phytotherapy with AI-driven personalization.
• Built the AI engine connecting questionnaire answers (symptoms, goals, age, species) to a phytotherapy-aware recommendation system selecting and combining ingredients from a curated herbal database.
• Designed safety constraints specific to natural remedies, including age limits (users as young as 3), species differences for pets, and known ingredient interactions, keeping recommendations both personalized and safe.
• Built recommendation logic that stays strictly within wellness guidance, never presenting as medical advice.
• Built subscription and recurring-cure infrastructure (7-day, 15-day, 1-month, 2-month plans) plus the e-commerce layer for the general wellness shop.
• Treated privacy as a first-class requirement throughout, given the sensitive health-related data users share.
• Architected the system to scale to new ingredients and wellness categories without manual reconfiguration.
Built an AI Care Coaching Platform delivering personalized digital care coaching, end-to-end, owning both the AI engine and the full-stack application it runs inside.
• Built a conversational coaching agent that interprets user check-ins, maintains context on each user's progress over time, and generates personalized guidance and nudges dynamically, rather than pulling from a fixed content library.
• Designed the prompt architecture and state handling so the coach's responses stay consistent and genuinely reflect what each user reported, avoiding generic advice.
• Built the backend architecture and APIs connecting user data, check-in history, and coaching state.
• Built the front-end experience end-to-end so the AI coaching engine, data layer, and interface work as one coherent system rather than separate pieces bolted together.
• Architected the platform to extend to new coaching domains without a system rebuild, keeping the coaching logic and product experience tightly aligned as it grows.
Built Surgoed, a real estate marketplace app for Suriname covering homes, plots, vacation rentals, and business properties, end-to-end including the AI-driven search and pricing layer.
• Built location-based property search, detailed listing pages with multi-currency pricing (EUR, USD, SRD), favorites and saved searches, and category-specific browsing across rentals, sales, plots, and business properties.
• Designed a smart search system going beyond exact filter matches to surface relevant listings based on user intent.
• Built a recommendation engine suggesting homes tailored to each user's browsing and saved-property behavior.
• Implemented a price estimation feature giving buyers and sellers a data-driven sense of a property's fair market value in an otherwise opaque market.
• Architected the backend so search, recommendations, and pricing stay in sync with live listing data as new properties are added, scaling to new listings and property types without rebuilding core logic.