Engineering Intelligence · Built with Claude
Engineering intelligence, powered by Claude.
Hoang Thuyet is building an AI-native platform that turns engineering drawings, technical documentation and industrial machine data into actionable insight.
Our AI agents help engineers understand complex designs, investigate equipment anomalies and make informed decisions, with traceable evidence behind every answer.
- Running demos with prospective customers
- Grounded AI agents on Claude
- Engineers approve every action
- 2D drawing2D
- Engineering AIAI
- Feature understandingFeatures
- 3D representation3D
- Vibration
- 4.2 mm/s
- Temperature
- 67 °C
- Current
- 12.8 A
- RPM
- 1480
What we're building on Claude
Hoang Thuyet is building an AI-native engineering intelligence platform for industrial teams. In our prototype, Claude reasons over engineering drawings, technical documents and machine-condition data: it extracts specifications with citations and investigates sensor anomalies after deterministic signal tools compute features such as FFT peaks. Engineers keep final decisions. We are running demos with prospective customers.
- 1Engineering inputDrawings, manuals, sensor features
- 2Claude + toolsRetrieval, parsers, signal-analysis tools
- 3Structured analysisSpecifications, anomalies, likely causes
- 4EvidenceEvery claim cites its source
- 5Engineer reviewPeople decide what happens next
Why it matters
Machines stop. Knowledge is hard to find. The cost is real.
Industrial teams lose time and money when equipment fails unexpectedly and when the engineering knowledge to act sits in drawings and manuals no one can search. These public figures show the scale of the problem we work on.
- $1.4T
- lost each year to unplanned downtime by the world's 500 largest companies, about 11% of their revenueSiemens, The True Cost of Downtime 2024
- 27 h
- of unplanned downtime per month at an average large plant, down from 39 hours in 2019Siemens, The True Cost of Downtime 2024
- 53%
- of global electricity is used by electric motor systems (2023), rising to 72% in industryIEA 4E EMSA, Policy Brief #9 (2025)
- 24.3%
- of Vietnam's GDP comes from manufacturing (2024), the market where we startWorld Bank, manufacturing value added
Third-party figures, linked to their sources. They describe the industry, not Hoang Thuyet's results.
Platform
One intelligence layer for engineering and industrial operations.
Industrial information is fragmented across drawings, documents, machines, sensors and operational systems.
We are building an AI layer that helps connect these sources and turn them into engineering context that people and AI agents can use.
Engineering data
- 2D Drawings
- Technical Documents
- CAD Metadata
Industrial AI Platform
- Context EngineL1
- AI ReasoningL2
- Engineering AgentsL3
- Tools & Industrial SystemsL4
Machine data
- Vibration
- Temperature
- Current
- Speed
- Operational Signals
Context engine
One question about a machine pulls in the drawing, the manual, the sensor history and the maintenance record behind it.
The same bearing housing drawing used in Engineering AI is linked to the motor monitored in Machine Intelligence, so an answer can cite both.
Illustrative dataEngineering Intelligence
From engineering drawings to structured 3D understanding.
Engineering teams still spend significant time interpreting drawings, extracting specifications and rebuilding legacy designs.
We are developing AI-assisted workflows that understand technical drawings and convert engineering information into structured, machine-readable representations that can support downstream 3D and CAD workflows.
Drawing Intelligence
Understand more than geometry.
Drawing understanding
Identify dimensions, annotations, views and engineering features.
Specification extraction
Turn unstructured drawing information into structured engineering data.
Document context
Connect drawings with manuals, specifications and related engineering documents.
Engineering assistant
Ask questions about drawings and receive responses grounded in engineering source material.
The main bore is specified as Ø40 mm with a tolerance of ±0.02 mm.
Source
Bearing Housing Drawing · BH-120 Rev A
Section A-A, bore callout
Machine Intelligence
Turn sensor signals into machine understanding.
Industrial machines continuously generate signals.
We are building AI-assisted monitoring that combines sensor analysis, operating context and engineering knowledge to help teams identify abnormal behavior and investigate potential equipment issues.
Sensor Pipeline
From raw signals to actionable context.
Layer 1 · Signal analytics
Deterministic processing produces measurable features.
Machine
- Motor
- Pump
- Fan
- Gearbox
Sensors
- Vibration
- Temperature
- Current
- RPM
- Acoustic
Signal processing
- Filtering
- FFT
- Feature Extraction
- Trend Analysis
Anomaly detection
- Baseline comparison
- Drift & threshold events
Layer 2 · AI reasoning
Language models reason over features and context, not raw waveforms.
AI context engine
- Features + events
- Machine history
- Manuals
- Engineering knowledge
- Maintenance context
Industrial AI agent
- Grounded explanation
- Suggested checks
- Draft tasks
Operator / Engineer
- Reviews
- Decides
- Approves
Signal processing and anomaly detection produce measurable features: RMS levels, spectral peaks, trends and threshold events, computed with established signal methods.
AI reasoning combines those features with machine history, manuals, engineering knowledge and maintenance context to help engineers investigate problems. A language model is never asked to read raw vibration signals on its own.
Industrial AI Agent
An AI agent that understands the machine, not just the conversation.
The agent plans a multi-step investigation, calls controlled tools for sensor features, documents and maintenance records, and answers with sources. Any action it proposes waits for a person to approve it.
Agent activity
- Retrieve vibration history
sensor_history(M-04, vibration, 24h) - Analyze extracted signal features
signal_features(M-04) - Check temperature trend
sensor_history(M-04, temperature, 24h) - Retrieve motor manual
docs.search("M-04 manual bearings") - Review maintenance history
maintenance.records(M-04) - Compare operating context
context.operating_state(M-04)
The agent does not control machinery. It reads approved data sources and proposes actions for people to approve.
Agentic Workflows
Specialist agents, one orchestrated workflow.
Industrial questions cross disciplines. Our architecture routes each step to a specialist agent with its own tools and scoped context, while an orchestrator plans the workflow, checks the results and asks a person to approve anything that changes a system.
Orchestrator
Plan · route · verify · request approval
Drawing Agent
Reads drawings and specifications
- drawing.parse
- spec.extract
- docs.search
Signal Agent
Interprets extracted sensor features
- signal_features
- sensor_history
- baseline.compare
Maintenance Agent
Prepares inspections and work orders
- maintenance.records
- task.draft
- checklist.build
Knowledge Agent
Grounds answers in manuals and history
- rag.retrieve
- cite.sources
- memory.scope
Shared: grounding, scoped memory, evaluation traces, permission-aware tool access.
Architecture in developmentAI Architecture
Grounded AI for industrial workflows.
Retrieval and context engineering decide what the model sees. Claude reasons over that context, agents orchestrate tools, and people approve what happens next. Memory, grounding, evaluation and observability run across every layer.
- Scoped Memory
- Grounding
- Evaluations
- Observability
Inputs
- Engineering Drawings
- Machine Manuals
- Sensor Features
- Maintenance Records
- Operational Data
Retrieval + Context Engineering
Select, rank and assemble the right sources
Claude
Reasoning layer
Agent Orchestration
Plan multi-step work across specialist agents
Tool Use
- Engineering Tools
- Maintenance Systems
- Industrial APIs
- Business Systems
Human Approval
Required before important actions
Action
Executed through existing systems, fully traced
Built with Claude
Building industrial intelligence with Claude.
Claude serves as a reasoning layer within our Industrial AI architecture.
We are exploring how Claude can combine retrieved engineering knowledge, machine context and controlled tool access to support complex industrial workflows.
Context engineering
Build relevant context from machines, drawings, users and workflows.
RAG
Retrieve relevant technical documentation and engineering knowledge.
Grounding
Tie AI responses back to source documents and machine information.
Agentic workflows
Coordinate multi-step engineering and maintenance workflows.
Tool use
Allow agents to interact with controlled engineering and operational tools.
Scoped memory
Maintain appropriate machine and workflow context across interactions.
Evaluation
Test groundedness, citations and tool choices before workflows ship.
Observability
Trace every retrieval, tool call and decision for review.
AI-native from day one
We build the platform itself with AI. Our engineering work runs on Claude Code agents for implementation, review and testing, under written guardrails: one change per pull request, tests that can fail, and a human decision at every open design question.
Evaluation
Industrial AI needs to be measurable.
Before AI-assisted workflows can be trusted in industrial environments, their behavior needs to be evaluated systematically.
| Dimension | Status |
|---|---|
| Groundedness | Testing |
| Citation quality | Testing |
| Tool selection | Testing |
| Context retrieval | Testing |
| Task completion | Testing |
Test Cases
Questions and tasks with expected sources and tools
Evaluation Runs
Repeatable runs against each model and prompt version
Agent Traces
Step-by-step record of planning and decisions
Tool Calls
Every call, its arguments and its result
Source References
Which documents and signals supported each answer
- plan
Investigate abnormal vibration on M-04 - tool
sensor_history(M-04, vibration, 24h) - tool
signal_features(M-04) - tool
docs.search("M-04 manual bearings") - source
Motor M-04 Manual · §6 Bearings - check
Every claim cites a source
Use cases
Built for engineering-intensive operations.
Engineering digitization
Transform legacy drawings and engineering documentation into structured digital knowledge.
Equipment intelligence
Understand machine behavior using sensor signals and operating context.
Maintenance assistance
Help maintenance teams investigate anomalies and prepare inspection workflows.
Industrial knowledge
Give engineers grounded access to manuals, drawings, specifications and machine history.
Safety & human control
AI assistance with engineers in control.
Grounded outputs
Connect AI responses to engineering sources.
Human approval
Require approval before important actions are executed.
Traceable reasoning
Record retrieval, tool calls and workflow execution.
Permission-aware access
Only provide agents with authorized context and tools.
The platform is designed as engineering decision support. AI-generated analysis should be reviewed by qualified personnel before safety-critical decisions are made.
Why Hoang Thuyet
Industrial AI that engineers can check.
When anyone can build an AI demo quickly, what matters is whether an engineer will trust the output on a real machine. Our product decisions start from that question.
- 01
Built on the data others skip
Technical drawings and machine signals are hard inputs: geometry, tolerances, spectra, trends. We design for them first, instead of wrapping a chat window around documents.
- 02
Signals measured, then reasoned about
Signal processing computes features with established methods. Claude reasons over those features plus manuals and history. Each layer does the job it is reliable at.
- 03
Every answer shows its source
Responses cite the drawing section, manual page or sensor history behind them, so an engineer can check the claim in seconds.
- 04
Measured before it is trusted
Groundedness, citations and tool choice are tested in an evaluation suite designed to rerun whenever a new Claude model ships.
Stage & roadmap
Demo stage, with a clear path to pilots.
We are demoing to prospective customers and shaping the first pilot around their drawings and machines.
- Demo
- Product demos running with prospective customers
- 2
- Core capabilities: Engineering AI and Machine Intelligence
- 4
- Specialist agents in the architecture, under one orchestrator
- 100%
- Of actions that change a system wait for human approval
The problem
Engineering knowledge is locked in drawings, manuals and machine signals that people read by hand. Interpreting a legacy drawing or an unusual vibration trend still takes an experienced engineer hours.
Why now
Models can now read technical documents and use tools reliably when they are grounded in sources and checked by evaluations. That makes decision support for engineers practical, not just impressive.
Business model
Software subscription for engineering and maintenance teams in manufacturing. Pricing is being validated in our customer demos.
Now
- Customer demos
- Drawing understanding with cited answers
- Sensor anomaly explanations grounded in manuals
Next
- First pilot with a demo customer
- AI-assisted 3D reconstruction workflow
- Evaluation results for each capability
- Hoang Thuyet connector for Claude (MCP)
Later
- CAD and maintenance system integrations
- More machine types beyond motors
- Listing on the Claude Marketplace
Company
Building practical AI for industry.
Hoang Thuyet is an early-stage team building engineering intelligence for industrial teams. The name comes from its founder, Hoang Cong Thuyet.
We are developing systems that combine engineering data, machine signals and AI reasoning to help industrial teams understand complex information and improve engineering workflows.
- Company
- Công ty TNHH Hoàng Thuyết · Tax code 0306037857 · Ho Chi Minh City, Vietnam
- Stage
- Demo stage · running demos with prospective customers
- Focus
- Industrial AI · Engineering Intelligence · Machine Intelligence
Founder
Hoang Cong Thuyet
Founder
Engineer and product builder with a decade in factory automation at manufacturers including Canon and Samsung, now focused on AI for engineering data and industrial equipment.
Contact
Build the next generation of industrial intelligence.
Interested in Industrial AI, engineering automation, machine intelligence or early product access? Let's talk.