You are hereIntro1/7
Open to senior AI & software engineering work

Hassan Javaid Rao

AI Engineer · Senior Software Engineer

LLM Applications, Multi-Agent Systems & Voice AI

I build applied LLM systems that survive production — multi-agent pipelines, RAG assistants and real-time voice agents — on top of five years of multi-tenant SaaS and AWS architecture.

Hassan Rao
Years of professional experience
5+Years of professional experience
Orders billed identically to human entry
71.7%Orders billed identically to human entry
Cost per order after optimisation
$0.008Cost per order after optimisation
Five-star client reviews
250+Five-star client reviews
About

I take LLM features from concept to production

I'm an AI-focused Senior Software Engineer with 5+ years of professional experience, including four-plus years building GCC enterprise logistics systems that process thousands of shipments daily across Oman, the UAE and Saudi Arabia.

I initiated and lead the AI programme at one of Oman's largest logistics platforms — shipping RAG chatbots, AI copilots, and a production multi-agent document-AI system that replaced most of a 12-hour daily manual data-entry workload. I proposed it, architected it, and took it to production.

Independently, I built and run AI Caller, a multi-tenant AI call-agent SaaS that answers live phone calls with hallucination-resistant, knowledge-grounded responses. Solo build, from the Twilio bridge through to billing and onboarding.

What I've learned from both: the model is the easy part. Grounding, evaluation and failure taxonomies are what separate a demo from something a business can bill against. I'm also a Section Leader at Stanford's Code in Place programme, and I mentor the engineers on my team through code review and architectural guidance.

At a glance

  • 5+ years of professional software engineering
  • 4+ years building GCC enterprise logistics systems
  • Applied LLM engineering: multi-agent, RAG, voice AI
  • Multi-tenant SaaS and scalable AWS architecture
  • Section Leader, Stanford Code in Place
  • 250+ five-star reviews as a Top Rated freelancer
Technical skills

What I build with

The stack behind the work below, grouped the way I actually use it.

AI / LLM engineering

  • Multi-agent orchestration (LangGraph)
  • RAG pipelines
  • Embeddings & vector search (pgvector)
  • Vision & document AI
  • Prompt engineering & grounding
  • Constrained decoding
  • LLM function/tool calling
  • LLM-as-judge & offline evaluation
  • Hallucination mitigation
  • Voice agents (Deepgram, Twilio)
  • Google Gemini APIs

Backend

  • Python (FastAPI)
  • PHP (Laravel)
  • Node.js (NestJS)
  • TypeScript

Frontend

  • Vue.js
  • React.js / Next.js
  • Inertia.js
  • Tailwind CSS

Cloud & DevOps

  • AWS (Lambda, SQS, S3, EC2, CloudWatch)
  • Docker
  • GitHub Actions CI/CD
  • nginx
  • Redis

Databases

  • PostgreSQL (+ pgvector)
  • MySQL
  • MongoDB
  • Redis

Architecture

  • Multi-tenant SaaS
  • Modular monolith & microservices
  • Event-driven systems
  • REST API design
  • OAuth 2.0 / JWT

Quality & monitoring

  • Sentry
  • AWS CloudWatch
  • PHPUnit
  • Git
  • Agile/Scrum
Selected work

Projects & case studies

Three builds worth reading in detail — a multi-agent document-AI pipeline in production, an AI voice-agent SaaS I built solo, and a piece of pure backend engineering.

Waybill Data-Entry Agents — Multi-Agent Document AI

Five agents reading handwritten waybills into structured orders

2024 — Present

A production multi-agent pipeline that reads handwritten bilingual Arabic/English shipping labels into structured orders, cutting a 12-hour daily data-entry workload to 2–3 hours of review. Built on a fixed LangGraph state graph with one agent per responsibility.

Daily data entry, now review only
12h → 2–3hDaily data entry, now review only
Orders billed identically to humans
71.7%Orders billed identically to humans
Cost and latency per order
$0.008 · 23sCost and latency per order
Python (FastAPI)LangGraphGoogle Gemini (Flash + Pro)PillowPostgreSQL + pgvectorLaravel+1

AI Caller — Multi-Tenant AI Voice-Agent SaaS

An AI receptionist that answers real inbound calls, 24/7

2025 — Present

A SaaS where any business self-serves an AI phone receptionist that answers real inbound calls in a natural voice — live in production on AWS. Founder and solo build, from the telephony bridge through to billing and onboarding.

Live inbound calls answered
24/7Live inbound calls answered
Self-serve signup and plans
Multi-tenantSelf-serve signup and plans
Telephony through to billing
Solo buildTelephony through to billing
Python (FastAPI)LaravelVue.jsPostgreSQL + pgvectorDeepgram Voice AgentTwilio+2

Translation Management API

Millions of records, streamed at low memory cost

2023

A production-grade REST API for locale-based translations: Controller → Service → Repository architecture, Redis tag-based cache invalidation, and cursor-based export streaming millions of records at low memory cost.

Records streamed per export
MillionsRecords streamed per export
Cache invalidation strategy
Tag-basedCache invalidation strategy
With a full PHPUnit suite
DockerizedWith a full PHPUnit suite
PHP 8.2LaravelMySQLRedisDockerPHPUnit
Experience

Where I've done it

Five years of continuous ownership, plus teaching and freelance work.

  1. Thiqa Tech

    Muscat, OmanCurrentDec 2021 — Present

    Promoted from Software Engineer to Senior Software Engineer. Five years of continuous ownership of the same multi-tenant logistics SaaS platform.

    Senior Software Engineer

    Jan 2024 — Present
    • Proposed, architected and lead the platform's AI programme — taking LLM features from concept to production for one of Oman's largest logistics companies, across 15 branches
    • Designed and shipped a five-agent document-AI pipeline (LangGraph + Gemini) that reads handwritten bilingual Arabic/English shipping labels into structured orders, cutting a 12-hour daily data-entry workload to 2–3 hours of review
    • Diagnosed the failure that killed the naive single-prompt version — correct digits assigned to the wrong field — and re-architected it as one agent per responsibility, making cross-field contamination structurally impossible
    • Built a shadow-mode evaluation harness: 2,700+ orders predicted alongside human operators and scored field-by-field, reaching 71.7% of orders billed identically to human entry
    • Engineered cost and latency down to $0.008 and 23s per order through model routing; one targeted fix cut expensive verification escalations from 78% to 20% of orders
    • Built RAG-based customer chatbots for shippers over tenant shipment data and policy documents, with grounding rules that keep answers scoped and hallucination-free
    • Shipped "Ask AI" for operations — a natural-language RAG assistant replacing minutes of manual cross-system lookups with instant grounded answers
    • Architected and led the migration from a single monolith to a modular architecture, driving incremental extraction with zero downtime
    • Engineered a high-performance export pipeline on AWS Lambda + SQS exporting millions of records at minimal cost, and an internal real-time chat system used across branches (NestJS, MongoDB, Next.js)
    • Mentor junior engineers through code reviews and architectural guidance; maintain 99.9% uptime with Sentry and CloudWatch observability

    Software Engineer

    Dec 2021 — Dec 2023
    • Core engineer on a multi-tenant logistics SaaS built from zero, scaling it to thousands of shipments per day and millions of records across Oman, the UAE and Saudi Arabia
    • Designed the platform's role-based access control spanning tenants, branches and operational roles
    • Built secure REST APIs with OAuth 2.0 — Password Grant for first-party mobile apps, Client Credentials for machine-to-machine integrations with DHL, AJEX and iMile
    • Improved system performance by roughly 20% by refactoring legacy code into modern patterns, and introduced Redis-backed queues to keep the UX smooth under heavy load
    LaravelVue.jsNode.jsPythonMySQLRedisAWS
  2. Fiverr & Upwork

    2020 — 2021

    Freelance Full-Stack Developer

    2020 — 2021
    • Top Rated on both platforms with 250+ five-star reviews
    • Delivered web and automation solutions for 200+ clients across logistics, e-commerce and SaaS
  3. Stanford Code in Place

    2025

    Section Leader (Teaching)

    2025
    • Competitively selected to teach Python to international learners in Stanford's flagship open-access programme
    • Led weekly sessions and gave personalised feedback

Education

B.S. Software Engineering

Sukkur IBA University

2017 — 2020

  • Fully funded merit-cum-need scholarship (4 years)
  • IELTS 7.0
  • Coursework: AI, Data Structures & Algorithms, OOP, Database Systems, Operating Systems, Web Engineering

Awards

  • Winner — Web Design Competition, Sukkur IBA
  • Winner — Code in the Dark, Sukkur IBA
  • 4-year fully funded merit scholarship
Reviews

What clients say

Verified LinkedIn recommendations and Fiverr reviews, quoted in full.

We had the pleasure of working with Hassan Javaid Rao on a complex website migration and setup project, and we can't recommend him highly enough. We needed a full transfer of an existing website to a new account with a new domain, including installation of all identified plugins and applications.

Hassan handled every aspect of the project with exceptional skill and attention to detail. He meticulously transferred the entire site, ensured that all necessary plugins and applications were properly installed, and made the process seamless. His technical expertise was evident in every step, and his communication was outstanding. He kept me updated throughout and was always available to answer any questions.

If you're looking for a professional who is reliable, knowledgeable, and dedicated to delivering high-quality results, Hassan is the perfect choice. His work was flawless, and we couldn't be happier with the outcome. Five stars all the way!

Anton Skileski
Architect, Retired
November 7, 2024

I highly recommend Hassan Javaid Rao for any web development project. He created an exceptional website for me using PHP, HTML, and CSS that not only looks fantastic but also performs seamlessly across phones, tablets, and desktops. The site runs fast, and his attention to detail in both functionality and design is top-notch.

What truly impressed me was his dedication to making sure I was fully satisfied with the final product. He made several revisions until the website was exactly what I envisioned. Even after completing the project and receiving payment, Hassan went above and beyond by making additional tweaks to ensure everything was perfect. His professionalism, technical expertise, and commitment to client satisfaction make him an outstanding developer, and I would work with him again in a heartbeat.

Alex Lluch
Manager, Villa Fabulosa & Villa Magnifica, Temecula Wine Country
🇺🇸 United StatesSeptember 28, 2024

Hassan is very thorough and honorable. I would recommend him to complete any technology related task with efficiency!

Ryan Henderson
Founder, Kream & Sugar
August 23, 2024

Hassan J. truly knocked it out of the park with his exceptional website development skills! He delivered a bug-free and professional project that EXCEEDED expectations in no time. With his quick responsiveness and politeness, working with him was a breeze. I genuinely hope this is the start of a great partnership! Highly recommend his services.

samblack366
🇺🇸 United States1 year ago

this was mybsecond time working with hassa. it was amazing working with him. He was extremely patient and committed to the project. He offered great advise and made the whole experience comfortable. I will be working with him again

acrossmediaRepeat client
🇨🇦 Canada5 months ago

I had a great experience working with Hassan. He was patient, responsive, and open to feedback throughout the project. Although my requests were required some adjustments along the way, he handled everything with a positive attitude and professionalism. The result was what I needed. Highly recommend.

yukipin00
🇻🇳 Vietnam1 year ago

Quoted verbatim, including each client's own wording and spelling.

Contact

Let's talk about what you're building

Whether it's a senior role, a freelance project, or a second opinion on an AI architecture decision — I read everything and reply within one working day.

Find me on

Or send a message

I reply to everything within one working day.