Forward Deployed Engineer
Prototypes into production. AI systems that hold up.
Next.js · Node.js · LangChain · MCP · Voice AI · Real-estate ops
Anas Aqeel
Forward Deployed Engineer
The developer who works with founders to take prototypes into production.
AI prototype that demos well, falls over in production. I make it hold. Retries, timeouts, per-call traces, real integrations, the parts that decide whether it survives its first busy day.
You get the unglamorous parts, done right.
You get a production-grade system. Not a working prototype, not a handoff to your team to figure out. The whole path, owned: reliability, observability, real integrations, the retries and timeouts and trace logs that tell you what actually happened.
Recent work sits at the intersection of AI systems, reliability, and integrations: voice agents that stay up when the phones do not stop, valuation pipelines that carry a source URL for every number, multi-agent command centers where every human correction feeds back into the prompts.
You ship faster because someone owns the whole path. Small loops, real production, no hand-off.
Experience
Calltura
- Co-founded a voice-AI company building phone agents that answer, qualify, and route inbound calls for UK trades and clinics. Own the entire technical side.
- Architected the real-time voice stack end-to-end: LiveKit for media transport, Telnyx SIP for UK telephony, Gemini Live for the conversational model, Cal.com for live booking, Supabase for state.
- Built the multi-tenant backend and operator dashboard: per-client agent configuration, call transcripts, booking flows, and an MCP tool layer the agent calls at runtime.
- Work the non-engineering side too. Positioning, compliance (UK GDPR and PECR), pricing, and vendor economics down to per-minute call cost.
Next.js
TypeScript
PythonSupabase
OpenAI IntegrationFlipIQ
- Longest-running engagement. Four production systems in daily use at a US real-estate investment firm, with a fifth in build.
- Built OMS end to end, the client onboarding portal. Replaced a 46-task spreadsheet with a real state machine across three roles, where blocking logic auto-unlocks manager tasks the moment the client completes their step.
- Contributed to Command, the CRM the firm runs its deal flow on, as part of a multi-developer team. It surfaces hot deals, comparable properties, and investment analysis, alongside agent contacts, track records, and a performance leaderboard.
- Within Command, owned the front-end work and built the AI role-play feature end to end, letting agents rehearse deal conversations against a voice agent rather than on live calls.
- Built Command Center end to end, the AI layer above the CRM. One router dispatches to eight specialist agents, each with its own persona, memory, and rules, unifying sales, tasks, email, and calendar behind a single interface.
- Built ARV IQ, a valuation tool that takes any property and returns comparables, estimating After-Repair-Value and cost of repair through a multi-stage pipeline.
- Currently building the property pipeline platform: acquisition through resale in one system, with a Section-Group-Field data model and one audited write engine so humans and AI agents obey the same access rules.
- Picked up real deal-flow domain knowledge (property data, transaction mechanics, investor workflows) while shipping against tight deadlines.
Next.js
Node.jsTypeScript
LangChainOpenAI Integration
n8nProject Twelve
- Built AI-driven product features end-to-end across Next.js, Node.js, and TypeScript. Ambiguous spec through shipped production code, owned throughout.
- Designed API surfaces and data models connecting AI agent workflows to a user-facing product.
- Worked directly with the founding team on a fast-moving roadmap focused on practical AI applications for personal growth.
Next.js
ReactTypeScriptNode.jsOpenAI IntegrationWattif (Brackets.ai)
- Shipped production features for wattif.ai, a live customer-facing platform. Full stack: Next.js and Tailwind frontend, backend integrations, data flows.
- Translated Figma designs into responsive, accessible UI. Debugged across the stack to resolve product issues quickly.
- Worked with the founder and the Brackets.ai team across timezones to keep the product moving.
Next.jsReactTypeScript
Tailwind CSSNode.jsPanaverse
- Built reusable Next.js and Tailwind page templates for product launches. Production-ready hero, pricing, feature, and CTA blocks.
- Focused on responsive, conversion-oriented landing systems with a clean TypeScript codebase.
Next.jsTypeScriptTailwind CSSReactSix systems. Every one in production.
Each card opens the full case study on the blog. Named clients kept anonymous until sign-off.
Calltura is my own product. AI agents that answer, qualify, and route inbound calls for businesses. The hard part is sub-second turn latency plus interruption handling. Every stage has its own timeout, fallback, and full per-call trace.
Next.jsLiveKitOpenAI IntegrationNode.jsCalltura
AI command center unifying sales, ideas, tasks, email, and calendar behind one interface for a founder-led team. One router dispatches to eight specialist agents, each with its own persona, memory, and rules. Every human correction becomes a training snapshot with human-approved diffs and rollback.
Next.jsLangChainOpenAI IntegrationNode.jsCommand center for a real-estate CRM
Automated after-repair-value tool. Nine-stage pipeline pulls live listing data, finds comps, reads property photos with a multimodal model, and estimates repair costs from live local pricing. No stale tables, no year-built shortcuts. Every price carries its source URL.
Next.jsLangChainMultimodal AINode.jsReal-estate valuation tool
Client-onboarding portal that replaces a 46-task spreadsheet with a real state machine across three roles. Blocking logic auto-unlocks manager tasks the moment the client company completes their dependent step. Access control enforced at the query layer, not the UI.
Next.jsTypeScriptNode.js
PostgreSQLOnboarding management system
Full property lifecycle in one platform: acquisition, due diligence, escrow, construction, resale. Section-Group-Field data model with 29 sections and roughly 2,100 fields. One audited engine sits between every writer and the database, so humans and AI agents play by the same access rules.
Next.jsTypeScriptPostgreSQLNode.jsProperty pipeline platform
A personal system that researches, drafts, publishes, and analyzes social posts across X, LinkedIn, Instagram, and Facebook. Every daily brief and every draft is approvable by email in one click. Every correction feeds back into the prompts as an auditable diff.
LangChainOpenAI IntegrationNext.jsn8nAgent X
More shipped via direct contracts and long-running founder relationships.
Skills & technologies
The stack I build and ship production systems with, day in, day out.
Full-stack web
End-to-end product engineering across the modern stack.
Next.jsReactTypeScript
JavaScriptNode.jsTailwind CSS
REST APIs
GraphQL
Express.jsFrontend & UI
Translating designs into responsive, accessible interfaces.
Next.jsReactTailwind CSS
Material-UI
FigmaTypeScriptBackend & data
APIs, data models, and the systems that power the product.
Node.jsTypeScriptPython
MongoDBPostgreSQL
Firebase
Vector DatabasesREST APIsAI & agents
Agentic systems, RAG pipelines, and LLM-driven product features.
LangChainLangGraphMCPRAGOpenAI IntegrationCrewAIAutoGenVector DatabasesPrompt EngineeringMultimodal AIAutomation & workflows
Replacing manual work with auditable systems.
n8n
Zapier
Make.com
Microsoft Power AutomateAPI IntegrationTools & practice
Shipping fast without compromising quality.
Git
GitHub
Vercel
CursorSystem DesignDeployment PipelinesCertifications
IBM RAG and Agentic AI:Professional Certificate
IBMGoogle AI:Professional Certificate
Build AI Agents using MCP
IBMAgentic AI with LangGraph, CrewAI, AutoGen and BeeAI
Agentic AI with LangChain and LangGraph
Fundamentals of Building AI Agents
Build Multimodal Generative AI Applications
Advanced RAG with Vector Databases and Retrievers
Vector Databases for RAG: An Introduction
Build RAG Applications: Get Started
Certified Blockchain Developer
What clients say.
LinkedIn recommendations
“Anas, thank you for helping us out. It's been great to have you support our development team. We really appreciate your time and effort in helping us grow our Marketplace and the Command platform. Great job.”
Tony Diaz
Founder @ FlipIQ · 30+ Years, 1,100+ Flips
January 2026 · Tony was Anas's client
“Anas is an extraordinary developer. He is very quick at grasping the scope of the project. He has impressed me with his exceptional sense of detail and commitment to timelines.”
Have a prototype that needs to be production.
Forward Deployed Engineer, embedded with your team. Prototype to production, or the whole product from scratch. Voice AI, agentic systems, real-estate platforms, or wherever the hard problem lives.
I read everything myself and reply within a day.