Available for select collaborations

PK / UTC+05

Umar Abdullah / Senior Software + AI Engineer

Built for scale.Wired to think.

Ten years engineering scalable software. Five years shipping production AI—agents, retrieval, automation, and platforms built for real-world load.

Explore selected work Explore capabilities

                         o
                 .-============-.
            .-+*#@@@@@@@@@@@@@@@@#*+-.
         .=#@@@#*+=--::::::::--=+*#@@@#=.
       .*@@#+-.                      .-+#@@*.
      +@@*:           .-++++-.           :*@@+
  x  #@@-         .=#@@@@@@@@#=.         -@@#
    :@@#         +@@#+=----=+#@@+         #@@:
     #@@-         '=#@@@@@@@@#='         -@@#   o
      +@@*:           '-====-'           :*@@+
       .*@@#+-.                      .-+#@@*.
         '=#@@@#*+=--::::::::--=+*#@@@#='
            '-+*#@@@@@@@@@@@@@@@@#*+-'
                 '-============-'
                         x

Live spatial modelProduction intelligenceAgents / RAG / Automation

System active

10Years
software engineering

05Years
AI engineering

Operating rangeArchitecture → production → scale

Systems online
Open to complex work

A decade of engineering judgment

I design system that think.

YearsSoftware engineering

Deep foundations in architecture, full-stack products, distributed systems, automation, and technical leadership.

YearsAI engineering

Focused practice turning LLMs, agents, retrieval, and intelligent workflows into observable production systems.

Technology constellationProduction stack / live map

10Ysystem
depth

Next.jsReactTypeScriptTailwind CSSNode.jsExpressPythonREST / GraphQLLangChainOpenAI / GPT APIsMySQLRAG Pipelines

Architecture layer Intelligence layer17 capabilities mapped

Capability architecture

Senior depth across the entire system.

Not a collection of tools. A practiced ability to move from ambiguity to architecture to reliable output.

AI systems05 years

Intelligence that survives production.

Agentic workflows, retrieval systems, evaluations, guardrails, and observability designed as software, not demos.

Architecture10 years

Systems engineered for the next order of magnitude.

Event-driven backends, multi-tenant platforms, real-time data, and resilient services with clean operational boundaries.

Automation500K+ pages/day

Complex work reduced to dependable flows.

Distributed queues, browser automation, integrations, and recovery paths that turn fragile manual work into infrastructure.

Product30+ shipped

Technical depth with a product instinct.

From ambiguous brief to shipped interface: strategy, system design, implementation, instrumentation, and iteration.

LeadershipEnd-to-end

Clarity across code, teams, and decisions.

Architecture direction, pragmatic standards, mentoring, and communication that moves senior teams through complexity.

Frontend

Next.js / React / TypeScript / Tailwind CSS

Backend

Node.js / Express / Python / REST / GraphQL

AI / ML

LangChain / OpenAI / GPT APIs / RAG Pipelines / AI Agents

Infrastructure

MySQL / Docker / MongoDB / PostgreSQL / Bull / BullMQ

Selected systems

Proof over promises.

Six selected production systems. The full archive contains 111 reviewed projects, including 12 verified live deployments.

Architecture / intelligence / operations

System telemetry06 cases online

06production
systems

AI CustomerLive

JHB Property44

AI Visibility4

SaaS AnalyticsLive

AI / SaaS / Automation / KnowledgeScroll vector: horizontal

01 / 06Featured system

Architecture knowledge graph01 connected nodes

System coreAI Customer Support AgentEngineering System

Architecture stage Technology dependencyMove pointer to inspect depth

AI Customer Support Agent

LangChainNode.jsOpenAIMongoDBBullMQ

Open case study

02 / 06Featured system

Architecture knowledge graph07 connected nodes

System coreJHB Property IntelligenceDecision Intelligence

01 / AcquireMLS data enters as an auditable run

02 / LearnHistorical behavior becomes local market models

03 / ValueActive listings receive defensible comparables

04 / DecideThree engines separate facts, market, and deal fit

05 / ExplainAI enrichment stays behind deterministic controls

06 / OperateReview, feedback, and export close the loop

Architecture stage Technology dependencyMove pointer to inspect depth

JHB Property Intelligence

MLS ingestion, market-model training, three-stage deal scoring, AI enrichment, maps, watchlists, and export in one operating system.

Next.jsNode.jsExpressMySQLAnthropicOpenAIGoogle SheetsDocker

Open case study

03 / 06Featured system

Architecture knowledge graph07 connected nodes

System coreAI Visibility PlatformAI SaaS Platform

01 / WorkspaceBrand intent enters through dedicated product surfaces

02 / ResearchSearch evidence grounds every automation

03 / OrchestrateLong-running work moves through durable queues

04 / ReasonA shared model layer routes by task and provider

05 / PublishGenerated assets cross a controlled integration boundary

06 / MeasureTraditional and AI search signals return to the workspace

Architecture stage Technology dependencyMove pointer to inspect depth

AI Visibility Platform

Model routing, research, queues, content generation, AI visibility, reporting, and publishing coordinated as a production platform.

Next.jsNode.jsLangChainOpenAIAnthropicGeminiRedisMySQL

Open case study

04 / 06Featured system

Architecture knowledge graph01 connected nodes

System coreSaaS Analytics PlatformEngineering System

Architecture stage Technology dependencyMove pointer to inspect depth

SaaS Analytics Platform

Next.jsNode.jsPostgreSQLRedisStripe

Open case study

05 / 06Featured system

Architecture knowledge graph07 connected nodes

System coreGrounded AI ChatbotConversational AI

01 / ConfigureEach bot gets its own behavior and brand boundary

02 / IngestHeterogeneous sources become a common knowledge format

03 / IndexChunks become tenant-scoped semantic memory

04 / EvaluateThe graph decides when retrieval is worth using

05 / GenerateModels answer within the bot policy and retrieved context

06 / DeliverOne backend serves hosted chat and website embeds

Architecture stage Technology dependencyMove pointer to inspect depth

Grounded AI Chatbot

Document ingestion, retrieval grading, multi-model generation, conversation operations, billing, and a distributable chat widget.

Next.jsExpressLangChainLangGraphOpenAIAnthropicGeminiPinecone

Open case study

06 / 06Selected build

Architecture knowledge graph01 connected nodes

System coreWeb Scraping Automation EngineEngineering System

Architecture stage Technology dependencyMove pointer to inspect depth

Web Scraping Automation Engine

PuppeteerNode.jsBullMQDockerMongoDB

Open case study

There is more beneath the surface.

View the complete work archive

Career signal

Built through real constraints.

Each role added a new layer: delivery, scale, architecture, intelligence, and leadership.

10Y05Y AI

2021 – 2024Remote

Senior Full-Stack Engineer

Marketlytics

Built and scaled a multi-tenant SaaS platform from 0 to 50K monthly active users. Designed microservices architecture. Implemented Stripe billing, automated onboarding, and real-time dashboards.

2019 – 2021On-site

Software Engineer

E-Khata

Delivered 15+ client projects across web scraping, automation, and web apps. Built Puppeteer-based scraping infrastructure processing 200K pages/day. Developed RESTful APIs and integrated third-party services.

2015 – 2019On-campus

B.Sc. Computer Science

Virtual University

Graduated with honors — specialization in Software Engineering. Capstone project: AI-powered plagiarism detection system. Active member of the Computer Science Society.

Engineering notes