ASIL ABDULDAEM Contact
Tripoli, Libya — open to opportunities

AI and Digital Transformation Engineer

I build the systems that turn oil & gas operations into data — forecasting models, LLM agents, and pipelines that hold up in a domain where being wrong is expensive.

View selected work Download CV

PYTHON LANGGRAPH AWS dbt TERRAFORM POSTGRESQL

01 — About

I build the data and AI layer that operations actually run on.

At PETGO — Petro Engineering and Technical Group, an oil & gas services group in Libya — I've shipped production forecasting models, a multi-tenant LLM agent platform that lets engineers query live databases in Arabic and English without writing SQL, and the operational systems that replaced their paper workflows.

Alongside that I architect data platforms: a medallion lakehouse on AWS unifying four source systems into one governed warehouse, with dbt marts, CDC ingestion, Step Functions orchestration, and infrastructure defined entirely in code.

The through-line is the same either way — systems that stay correct when nobody is watching them. Tested, observable, reversible, and documented well enough that someone else can own them.

0Years in oil & gas
0Operational units live
0Source systems unified
0Marts under test
02 — Experience

Where I've been building.

PETGO — Petro Engineering and Technical Group

AI & Digital Transformation Engineer

Dec 2024 — Present
  • Built ML well-production forecasting on historical production, well-test results and reservoir properties, replacing manual estimates in planning and reporting.
  • Deployed ACE, an LLM agent platform with Arabic and English NLP, letting engineers query operational databases without SQL — with RAG and vector search over internal documents.
  • Developed voice AI agents for hands-free data retrieval and reporting in field environments.
  • Designed a digital work-permit platform that replaced paper approvals with live status updates and supervisor notifications.
  • Built an internal talent evaluation platform in Next.js, React and PostgreSQL, standardising technical assessment and candidate tracking.
  • Delivered Power BI executive dashboards across 5 operational units with automated refresh.

PETGO — Petro Engineering and Technical Group

Data Analyst — Oil & Gas

May 2024 — Nov 2024
  • Worked across the company's production and well-test data systems to monitor well performance and support engineering analysis.
  • Maintained and improved allocation logic and rules distributing production to individual wells and facilities.
  • Cleaned, validated and organised large volumes of production and well-test data underpinning reports and models.
  • Built Power BI reports surfacing trends, anomalies and underperforming wells for engineers and managers.

Presto — Food Delivery

Data Engineer

Present
  • Architected an AWS medallion lakehouse (S3 + Redshift) unifying four source systems — MySQL, Odoo/PostgreSQL, MongoDB and BigQuery — into one governed warehouse.
  • Eliminated all analytical load from the production database with Glue batch and DMS change-data-capture ingestion, repointing every BI tool onto Redshift.
  • Stopped ongoing permanent loss of clickstream history with a daily BigQuery → S3 archival pipeline and lifecycle tiering.
  • Built a dbt project from staging through marts with 100% test coverage, establishing one agreed definition per business metric.
  • Implemented Step Functions orchestration with retry logic and CloudWatch/SNS alerting, taking pipeline failure detection from 0% to 100% within a 15-minute window.
  • Moved all pipeline code into Git with Terraform/CloudFormation IaC and GitHub Actions CI/CD, replacing console-edited scripts.
03 — Selected work

Systems in production.

BUILT & SHIPPED

01 / FLAGSHIP

ACE — AI Chat Engine

A multi-tenant LLM agent platform for production engineers and analysts. A LangGraph agent queries live operational databases in natural language — strictly read-only, behind a query guard — across PostgreSQL, MySQL and SQL Server, in Arabic and English. Each department brings its own database and its own LLM provider. Adds document analysis, voice I/O, generated reports, and five-tier role-based access.

Removes SQL as a prerequisite for field data Two sellable editions on one codebase
LangGraphLangChainMCPFastAPI Anthropic / OpenAI / GeminiPostgreSQLQdrantRBAC
02

ML Well Production Forecasting

Predicts well output from historical production, well-test results and reservoir properties, feeding planning and reporting cycles instead of manual estimates.

Manual estimates → a model in the planning cycle
Pythonscikit-learnPostgreSQLPower BI
03 / FLAGSHIP

Presto — Lakehouse Migration

Presto, a food delivery company, was running analytics directly against its production database and deleting clickstream history every day to stay under a storage cap. I designed and delivered a medallion lakehouse — S3 Bronze, Redshift Silver/Gold — unifying MySQL, Odoo/PostgreSQL, MongoDB and BigQuery behind one governed warehouse, with dbt marts defining every business metric exactly once.

Analytical load on prod
ongoing→zero
Clickstream retained
deleted→100%
Failure detection
0%→100%
Marts under test
0%→100%
AWS GlueDMS (CDC)RedshiftS3 dbtStep FunctionsTerraformPySpark
04

PETGO Pulse

Intranet project and contribution tracker — accounts, performance cycles, project tracking and notifications, built on a tokenised design system with full light and dark themes and a test suite across both ends of the stack.

Consolidates HR and project reporting
Django 5DRFReactTanStack Query Tailwind v4PostgreSQLpytestVitest
05

Voice AI Field Agents

Hands-free voice interface for field operations. Spoken commands retrieve readings and log reports without a keyboard or a stable connection.

Built for gloves-on field conditions
PythonWhisperTTSn8nREST
06

Digital Work Permit Platform

Replaced paper-based permit-to-work approvals with a tracked digital workflow, including approval chains and an auditable history.

Approval time: days → hours
Next.jsPostgreSQLRESTDocker
07

Talent Evaluation Platform

Structured assessment tool for technical evaluation and candidate tracking, standardising how candidates and staff are scored across departments.

One rubric across all departments
Next.jsReactTypeScriptPostgreSQL
08

Executive BI Dashboards

Production, finance, and HR reporting for leadership — live data with automated refresh cycles replacing hand-assembled monthly decks.

Live across 5 operational units
Power BIDAXSQLETL
04 — Capabilities

What I work with.

AI Systems & Agent Engineering

  • Natural-language → SQL agents
  • Read-only query guardrails
  • LangGraph state machines
  • Agent memory & checkpointing
  • Model Context Protocol (MCP)
  • Tool-calling & function schemas
  • Provider-agnostic LLM abstraction
  • Multi-tenant agent isolation
  • RAG & embeddings
  • Vector search
  • LLM observability (LangSmith)
  • Prompt & persona systems
  • Voice agents (STT / TTS)
  • Arabic & English NLP

Data Platform & Modelling

  • Medallion architecture
  • Lakehouse design (S3 + MPP)
  • OLTP / OLAP workload separation
  • Change data capture (CDC)
  • dbt: staging → intermediate → marts
  • Dimensional modelling & grain
  • Surrogate keys & SCDs
  • Incremental models
  • Metric standardisation
  • Source-freshness SLAs
  • Reconciliation testing
  • Idempotent, replayable pipelines
  • Schema-drift handling
  • Redshift sort / dist keys

Cloud & Infrastructure

  • S3
  • Glue / PySpark
  • Redshift
  • DMS
  • Step Functions
  • Athena
  • CloudWatch / SNS
  • Terraform
  • CloudFormation
  • Docker
  • GitHub Actions
  • Linux / VPS

Governance & Security

  • Least-privilege IAM
  • Role-based access control
  • SSO federation (SAML)
  • Secrets management
  • PII anonymisation
  • Audit logging
  • Data catalog & lineage
  • Runbooks & DR procedures

Databases

  • PostgreSQL
  • MySQL
  • SQL Server
  • MongoDB
  • Redshift
  • BigQuery
  • Qdrant
  • Supabase

Backend & APIs

  • Python
  • SQL
  • FastAPI
  • Django / DRF
  • Flask
  • SQLAlchemy
  • Pydantic
  • Async & SSE streaming
  • C#

Frontend

  • TypeScript
  • React
  • Next.js
  • Tailwind
  • TanStack Query
  • Vite

Analytics & ML

  • Power BI / DAX
  • Metabase
  • Production forecasting
  • Feature engineering
  • Model evaluation
  • Jupyter

Engineering Practice

  • Test-covered services
  • CI gates on PR
  • Infrastructure as code
  • Trade-off analysis & ADRs
  • Phased migration roadmaps
  • Technical writing

Let's build something that holds up.

Open to roles and collaborations in data engineering, AI systems, and digital transformation — remote or based in Tripoli.

aseeldayem16@gmail.com Download CV
GitHub LinkedIn Tripoli, Libya