Morteza (Mort) Mogharrab, software engineer and systems architect in Calgary

I build production AI systems for engineering and industrial operations.

Right now that is Entrek Engineering's platform, built from an empty repository: it reads scanned documents, drawings and vendor bids, answers questions from project records with citations, and runs the operations portal around it, on AWS.

How the platform at Entrek fits togetherRead the case study
Process schematic: project documents (scanned PDFs, CAD drawings, spreadsheets and vendor bids) pass through OCR and LLM extraction, are validated and stored in PostgreSQL with vector embeddings, then feed three tools — cited answers using retrieval-augmented generation, a Text-to-SQL agent for data questions, and a P&ID copilot for drawing audits — delivered to engineers and managers through one portal.
Engineer
Morteza Mogharrab
Discipline
Software and AI systems
Status
Building at Entrek Engineering
Education
MSc Computing Science, U of Alberta, 4.0 GPA

What I build

I start from the operational problem — the re-keying, the searching, the spreadsheet nobody trusts — and own the software that removes it, end to end.

Operational platforms
Full-stack systems that replace spreadsheets and email threads: budgets and cost codes, procurement, field tickets, invoices and dashboards, behind single sign-on with role-based access.
Seen inEntrek platformPiko Digital
Document intelligence
Extraction pipelines that turn scanned PDFs, vendor proposals and engineering drawings into validated, structured records, with retries, schema checks and cost control.
Seen inEntrek platformExtraction pipeline
Retrieval and agents
Question answering over a company’s own records with citations, Text-to-SQL over its databases, and tool-using agents that carry out a workflow rather than just describe it.
Seen inEntrek platformAER compliance agent
Delivery to production
Architecture through deployment and upkeep: AWS, Docker, schema migrations, separate staging and production, and the guardrails that keep AI cost and accuracy predictable.
Seen inEntrek platform

Selected work

One enterprise platform in service, two open-source AI projects you can run yourself, and the platform work that came before.

Enterprise platformApr 2026 – presentIn service, ongoing

Entrek Engineering operations and AI platform

An oil and gas EPCM firm ran its operations on Microsoft 365 files, spreadsheets and email. I'm building the platform that replaces them, from an empty repository: AI extraction for scanned documents and vendor bids, a knowledge base with cited answers and Text-to-SQL, a P&ID copilot, automated design basis reports, industry intelligence, and the operations portal around them, deployed on AWS behind Microsoft SSO.

My role: Engineer responsible end to end: architecture, frontend, backend, AI, infrastructure.

  • React
  • TypeScript
  • FastAPI
  • PostgreSQL
  • RAG
  • Text-to-SQL
  • LLM agents
  • AWS
  • Azure AD
Bid leveling per package
2 days → <1 hrest.
First draft of a design basis memo
~2 days → ~2 hrsest.
Equipment list from a P&ID
~3 hrs → <10 minest.
Modules on one platform
17

est. Estimated from typical task durations before and after each workflow moved onto the platform; not a controlled measurement.

Agentic AI2025 – 2026Open source

AER compliance agent

A LangChain agent that runs an Alberta Energy Regulator calibration audit from one plain-English instruction: it searches the directives, checks equipment records, emails the report, schedules the follow-up and logs maintenance, using seven tools.

  • Python
  • LangChain
  • GPT-4o
  • ChromaDB
  • Pydantic
Audit on the sample facility
2 hrs → <30 sest.
Tools the agent chooses between
7

est. Manual baseline is an estimate.

Case studySource on GitHub

Document intelligence2025 – 2026Open source

Multi-model document extraction pipeline

A batch extraction pipeline that processes structured records concurrently through OpenAI, Anthropic or a local Ollama model, switched by configuration. A thread-safe rate limiter and four layers of output validation keep it within API limits and stop malformed results.

  • Python
  • OpenAI
  • Anthropic
  • Ollama
  • Pydantic
Records per minute
450+
First-pass parse accuracy, by backend
98.1–99.5%

Case studySource on GitHub

Web and mobile platformMay 2024 – Aug 2025Completed

Piko Digital platform

Admin and analytics platform plus a cross-platform React Native app for a social platform. I built the analytics dashboards, JWT and role-based access, moderation tooling and a shared component library for iOS, Android and web.

  • Spring Boot
  • React
  • React Native
  • PostgreSQL
  • GraphQL
Daily active users
1,000+
Less admin processing time
60%

Case studyConfidential, no source

Experience

  1. Apr 2026 – Present

    Calgary, Alberta

    Permanent Full-time

    Current role

    Software Systems & AI Developer

    Entrek Engineering Ltd.

    Building the firm's AI-integrated engineering and operations platform from scratch, replacing file- and spreadsheet-based Microsoft 365 workflows with one cloud-deployed portal.

    • Enterprise AI Knowledge Base & Text-to-SQL. Multi-modal ingestion (CAD, Excel, PPTX, PDF, Vision) with graph-based retrieval, LLM reranking and cited chat, plus a Text-to-SQL engine for multi-hop reasoning across relational and unstructured data. Estimated to cut document look-up from ~20 minutes to under a minute per question.
    • Industry Intelligence Platform. Monitors clients, competitors and projects via LLM web research and video summaries; delivers de-duplicated findings, executive briefings, an opportunity pipeline, trend detection, an entity directory and a RAG assistant with hybrid full-text/vector search. In rollout; designed to replace an estimated ~6 hours a week of manual market monitoring.
    • P&ID Intelligence Copilot. Deterministic DXF parsing and topology mapping combined with LLM agents to automate equipment extraction and process safety audits, reducing equipment-list extraction from an estimated ~3 hours to under 10 minutes per drawing.
    • Automated Engineering Reports (DBM). AI rewriting engine for Design Basis Memoranda plus an interactive Report Studio to generate, revise and export (DOCX/PDF) technical reports; first drafts in an estimated ~2 hours instead of ~2 days.
    • Procurement & Bid Automation. Parses unstructured vendor proposals, extracts technical and commercial data, and automates bid-spread leveling, comparison and validation, cutting bid leveling from an estimated ~2 days of Excel work to under an hour per package.
    • Operations Platform. Tracks budgets, cost codes, field tickets, purchase orders, invoices and meeting minutes with real-time burn charts, replacing an estimated day of spreadsheet consolidation with live dashboards.
    • Security & Cloud. Microsoft SSO (Azure AD) with feature- and document-level RBAC; AWS deployment on Docker, EC2, RDS and S3 with separated staging and production environments.
    • Python
    • FastAPI
    • PostgreSQL
    • SQLAlchemy
    • React
    • TypeScript
    • RAG
    • LLM Agents
    • Text-to-SQL
    • AWS
    • Docker
  2. Sep 2025 – Apr 2026

    Edmonton, Alberta

    Self-employed

    Independent Software & AI Developer

    Independent Practice

    Applied R&D period building domain-specific AI systems for Alberta's industrial sector.

    • AER Compliance Agent. LangChain agentic system orchestrating 7 tools (ChromaDB RAG, equipment DB, email, calendar, maintenance logging) to run autonomous compliance audits. Cut an audit of the sample facility from an estimated 2 hours to under 30 seconds.
    • LLM Document Intelligence Pipeline. Hybrid inference engine processing 1,000+ structured records concurrently across OpenAI, Anthropic and local Ollama at 450+ records/minute with thread-safe rate limiting.
    • Marketing Automation. n8n pipeline integrating Gemini and GPT-4 Vision, producing 200+ assets/day at a 97.5% cost reduction versus manual production.
    • Agentic AI
    • LangChain
    • RAG
    • Python
    • ChromaDB
    • n8n
  3. May 2024 – Aug 2025

    Edmonton, Alberta

    Contract Full-time

    Full Stack Software Engineer (Core Platform)

    Piko Digital Ltd.

    Worked across the full stack of a core social platform: microservices, frontend, mobile, data layer and operations.

    • Scale. Helped scale the platform to 1,000+ daily active users through microservice architecture and full-stack engineering.
    • Analytics Dashboard. Built a real-time analytics dashboard aggregating 15+ data sources, a 300% improvement in analytics.
    • Auth & Governance. Automated JWT auth and RBAC, cutting admin processing time by 60%.
    • Component Library. 50+ React Native/TypeScript components for iOS, Android and web; 40% faster development.
    • Mentorship. Mentored 3 interns and wrote 20+ SOPs; 40% faster onboarding.
    • Spring Boot
    • React
    • React Native
    • TypeScript
    • PostgreSQL
    • GraphQL
    • Docker

About

I'm Mort. I take an operational bottleneck that costs a team real time, learn the business logic behind it, and ship the system that removes it — then keep it running.

At Entrek Engineering I'm building the firm's whole platform: architecture, frontend, backend, AI pipelines, database and cloud. Before that I built domain-specific AI agents as an independent developer, and worked across the full stack of a social platform at Piko Digital as it grew past 1,000 daily users.

I'm comfortable owning the whole lifecycle, from first commit to a deployed system with real users, and I don't need perfect requirements to start.

Outside software, I've spent years as a community music director and have played piano and synthesizer since my teens. In 2025 the University of Alberta recognised my volunteer work with elderly and disabled communities with its Graduate Community Engagement Award.

Skills

Backend and data
Python, FastAPI, SQLAlchemy, Alembic, PostgreSQL, Java Spring Boot, REST, GraphQL
Frontend and mobile
React, TypeScript, Tailwind, React Native
AI and retrieval
LangChain and LangGraph, RAG pipelines, Text-to-SQL, vector search (Weaviate, Qdrant, ChromaDB), OpenAI, Anthropic, Perplexity, Ollama, Tesseract OCR
Infrastructure
AWS (VPC, EC2, ELB, RDS, S3, Route 53), Docker, Nginx, CI/CD, Azure AD single sign-on

Education and certification

  • Master of Computing Science, Multimedia

    University of Alberta, Sep 2023 – Apr 2025. GPA 4.0 / 4.0.

    Graduate Course-based Master's Community Engagement Award, Mar 2025

  • Mitacs Innovation Leader

    Mitacs, issued Nov 2026

If a workflow is costing your team days, I'd like to hear about it.

I'm based in Calgary and always open to conversations about hard problems, contract work, or what comes next — especially with engineering firms, environmental consultancies and industrial operators in Alberta.