Available for Projects · Top Rated · $95/hr

Industrial AI
Engineer

20+ years operating oil & gas plants. Now I build the AI that runs them.Predictive maintenance, operations intelligence, and AI automation — from someone who's actually worked the plant floor.

"I speak both P&ID and Python. I've calibrated transmitters and trained ML models. That combination is rare — and it's exactly what you need."
20+
Years Industrial Ops
3
SaaS Products Built
$95
Hourly Rate USD
100%
Job Success Score

// PROFILE

Profile & Credentials

Industrial AI engineering expertise — 20+ years operating plants, now building the systems that run them.

Freelance Stats

Client Success Rate100%
Rating★★★★★ 5.0
Response TimeWithin 2 hours
Time ZonePHT (UTC+8)
LanguagesEnglish (Fluent), Filipino
$95 / hour

Domain Expertise

Oil & Gas OperationsRotating EquipmentPredictive MaintenanceE&I EngineeringSCADA / DCSInstrument CalibrationP&ID AnalysisOEE / Plant KPIsOffshore PlatformsPETRONAS Standards

AI & Data Stack

PythonClaude APILangGraphIsolation ForestNext.jsSupabasePower BIRechartsMake.comPostgreSQLTime-Series MLETL Pipelines

Certifications & Talks

🎓 Data Science MastersSimplilearn
🎤 PyCon 2019 APACSpeaker
🏭 E&I Engineer20+ yrs
🌐 Platformhayahai.com
💻 GitHubargeoalecha

Bio

Before I built AI systems, I operated them. For over two decades I worked as an Electrical & Instrumentation Engineer across oil & gas operations — including offshore platforms and PETRONAS-standard facilities — calibrating sensors, troubleshooting rotating equipment, and keeping critical systems alive under pressure.

Today I translate that industrial depth into AI-powered systems: predictive maintenance SaaS, operations intelligence dashboards, and automated workflows that speak the language of plant floor data. I understand why your vibration sensor drifts, what a bearing wear signature looks like in the frequency domain, and how to build the ML pipeline that catches it before your compressor fails.

If you need someone who can read a P&ID and write a Python ML pipeline — I'm the rare hire you're looking for.


// WORK SAMPLES

Portfolio

Sample UI previews of deliverable types — real systems built for industrial + SMB clients.

ROTABRAIN · Pump P-101 Monitor⚠ ANOMALY
VIB-X (mm/s)
8.7
VIB-Y (mm/s)
6.2
TEMP (°C)
74.1
PRESS (kPa)
450
FLOW (m³/h)
38.2
⚠ Isolation Forest Score: -0.312 · Bearing wear signature detected · Recommend inspection within 48h

RotaBrain — Predictive Maintenance SaaS

Real-time anomaly detection for rotating equipment. Isolation Forest ML model ingesting vibration, temperature, and pressure sensor data. Flags bearing wear 48–72h before failure.

PythonIsolation ForestNext.jsSupabaseClaude API
OEE
84.2%
MTBF
720h
MTTR
4.2h
AVAIL
97.1%
Downtime by Loss Category (hrs/month)
Compressor K-201: Vibration threshold exceeded
Pump P-101: All parameters nominal

Plant Operations Intelligence Dashboard

OEE, MTBF, MTTR, and availability tracking. Downtime Pareto analysis by loss category. Real-time equipment health alerts. Built in Next.js + Recharts with Supabase backend.

Next.jsRechartsSupabasePower BIOEE
AGENTIC WORKFLOW · Maintenance Report Generator
📡SCADA Data
🤖Claude API
🔍RCA Agent
📄SOP Output
09:14:02→ Ingesting 72h sensor history for K-201
09:14:04⚠ Anomaly cluster detected at T-06:22
09:14:06→ Querying equipment manual (RAG)
09:14:09✓ RCA draft: Seal degradation likely cause
09:14:11✓ Maintenance SOP generated · 6 steps

AI Maintenance Report & RCA Generator

Multi-agent system using Claude API + LangGraph. Ingests SCADA/historian data, runs anomaly detection, queries equipment manuals via RAG, generates RCA and maintenance SOPs automatically.

LangGraphClaude APIRAGMake.comPython
DATA PIPELINE · Industrial ETL + Quality Scoring
INGESTED
48,291
rows
CLEANED
47,984
rows
FLAGGED
307
anomalies
QUALITY
98.7%
score
ETL-01Extract → OSIsoft PI Historian (API)
ETL-02Transform → Pivot, resample 1-min → 15-min
ETL-03Flag → 307 out-of-range sensor readings
ETL-04Load → Supabase PostgreSQL ✓
ETL-05Quality Score: 98.7% · Report sent ✓

Industrial Data ETL Pipeline

Python ETL ingesting from OSIsoft PI Historian, SCADA exports, and CSV sensor logs. Pivot/unpivot, resampling, outlier flagging, data quality scoring. Loads to Supabase PostgreSQL.

PythonPostgreSQLSupabasePower QueryETL

// SAMPLE PROPOSALS

Proposal Templates

Proven opening hooks for different engagement types. Lead with domain credibility, follow with technical capability.

Predictive Maintenance$2,000–5,000 fixed

Anomaly Detection for Industrial Sensor Data


"Before I built anomaly detection models, I calibrated the sensors they run on.

20+ years as an E&I Engineer in oil & gas — PETRONAS facilities, offshore platforms, rotating equipment — means I understand your data before I model it. I know what bearing wear looks like in raw vibration data. I know why your temperature sensor drifts after a process upset. I know the difference between a real anomaly and instrument noise.

For this project, I'll deliver: (1) Python Isolation Forest / LSTM pipeline tuned to your equipment type, (2) anomaly scoring with confidence bands, (3) a Next.js dashboard to visualize sensor health in real-time.

Can we schedule a 20-minute call to walk through your historian data structure?"

Operations Dashboard$1,500–3,500 fixed

Plant KPI Dashboard — Power BI / Next.js


"I've stood in control rooms watching operators struggle to get answers from fragmented systems. That's why I build dashboards differently — starting from what operators actually need to see, not what the data happens to export.

My background: 20+ years E&I in oil & gas, now full-stack AI/data engineer. I've built OEE trackers, MTBF/MTTR monitors, downtime Pareto dashboards, and real-time equipment health displays — all connected to real plant historian and SCADA data.

I'll deliver a production-ready dashboard in Power BI or Next.js + Recharts, connected to your data source, with drill-down capability and mobile responsiveness. Timeline: 3–4 weeks."

AI Automation$800–2,000 fixed

AI-Powered Maintenance Report Generation


"Your maintenance engineers shouldn't be writing reports — they should be fixing equipment.

I build AI agents using Claude API + LangGraph that ingest sensor history, run root cause analysis, query equipment manuals via RAG, and generate structured maintenance reports and SOPs — automatically.

I've done this in the industrial context specifically: I understand FMEA logic, equipment failure modes, and what a real RCA needs to contain. I'm not just an AI developer applying generic patterns to your problem — I've been the maintenance engineer on the other end of these reports.

Happy to show you a live demo of the prototype I've already built for rotating equipment."


// CLIENT EXPECTATIONS

What Industrial Clients Ask — And How I Answer

Common concerns from plant managers, reliability engineers, and ops directors. Addressed directly.

Domain-first, not framework-first
I've hired AI engineers before who needed weeks just to understand our process.

20+ years operating oil & gas facilities before writing a single ML pipeline. You won't spend the first month explaining what a historian is, why bearing temperatures trend before failure, or what a compressor surge signature looks like.

Deployed systems, not demos
I need something I can show operations management — not a Jupyter notebook.

Every engagement ships a running production system — a live dashboard, an integrated pipeline, or a deployed agent. Prototypes don't count as deliveries here.

Tuned against real process limits
How do I know the anomaly detection won't flood my team with false alarms?

Models are validated against your actual operating envelopes and alarm history — not generic thresholds. Domain knowledge means knowing which deviations are process noise and which ones matter.

Fixed scope. Fixed delivery.
Every IT project I've been part of went over time and over budget.

Engagements start with a scoped proposal — defined deliverables, defined timeline, defined price. If scope needs to change, we discuss it openly before any new work begins.


// PACKAGES

Service Packages

Fixed-scope engagements with clear deliverables. All packages include documentation and a 30-day support window.

Audit & Roadmap
$500fixed

AI readiness assessment + prioritized recommendations. Delivered as a structured report with quick-win opportunities and ROI estimates.

  • Process & data audit (2–3 workshops)
  • AI opportunity mapping (3–5 use cases)
  • Tech stack recommendation
  • Prioritized roadmap document
  • 1-week turnaround
Most Popular
Anomaly Detection Build
$2,500fixed

End-to-end predictive maintenance system for rotating equipment. From raw sensor data to live anomaly alerts with a dashboard.

  • Python ML pipeline (Isolation Forest / LSTM)
  • Sensor data ingestion + ETL
  • Anomaly scoring + threshold calibration
  • Next.js dashboard with live alerts
  • Supabase backend + API
  • Documentation + handover
  • 30-day support
Operations Intelligence
$3,500fixed

Full plant KPI dashboard — OEE, MTBF, downtime Pareto, equipment health — connected to your data source.

  • Data source integration (PI, SCADA, CSV)
  • ETL pipeline + Supabase storage
  • Next.js or Power BI dashboard
  • OEE, MTBF, MTTR, Availability
  • Downtime Pareto + drill-down
  • Mobile responsive
  • 30-day support
AI Agent Build
$4,500fixed

Custom LangGraph multi-agent system for maintenance report generation, RCA analysis, or compliance automation.

  • LangGraph multi-agent architecture
  • Claude API integration + RAG
  • Equipment manual knowledge base
  • Automated SOP / RCA generation
  • Make.com workflow integration
  • Full API + Next.js UI
  • 60-day support