Smart Factory
Bizarim AI

From welding quality prediction to AI that connects manufacturing work.
Explore the technologies and projects developed on the factory floor.

Explore related projects

Manufacturing AI technology history

Select “Related project” on a card to view its illustration below.

Open SVG
2018 to 2026
100%
Bizarim manufacturing AI technology history AI that used to be built separately for each area now connects inside agents from 2025. Anomaly and object detection still run as dedicated models. Vision OCR RPA ML Agents AI data Foundation Blended colour: technologies merge Right side TO-BE: research and direction Before AI 2018 to 2024 domain AI: built and trained area by area 2025 to 2026 connected inside agents TO-BE: the whole smart factory context, AI driven 2015 ~ 2017 2018 2019 2020 2021 2022 2023 2024 2025 2026 Technology groups Integration Vision Inspection, grading, 3D OCR Read displays and documents RPA Automate collection and entry ML Equipment data, frequency analysis, prediction Work agents Chat and task execution AI data Data supply and training datasets Foundation MES, costing, certification, devices View related project: PCB visual defect inspection AI solution pilots PCB defects, food packaging and expiry dates Reference images, PaDiM, YOLO, OCRRelated project View related project: Poultry carcass quality grading Poultry quality grading Shared encoder with grading heads DAB-DETR, YOLO, PatchCore, keypoints, active learningRelated project View related project: Video understanding and dedicated detection models Dedicated detectors Anomaly and object models called by VLM RF-DETR, DinomalyRelated project View related project: Exploring 3D reconstruction for weld inspection 3D reconstruction research (2022 ~) Internal research for weld inspection, leading to video monitoring and reading NeRF and Instant-NGP through to DUSt3RRelated project View related project: Video understanding and dedicated detection models Video understanding Safety and quality VLM, VGGT, Depth Anything Vision+OCRRelated project View related project: Pilot factory vision inspection Pilot factory vision Worker-friendly smart factory pilot YOLO, transfer learningRelated project View related project: Robotic vision for missing-part inspection Robotic vision R&D Preset camera positions, lighting control, missing-part inspection YOLO, reference images, PLC socketsRelated project View related project: Extending missing-part inspection to bus frames Bus-frame inspection Missing parts, scaled up SAM, feature matching, reference imagesRelated project View related project: Reading equipment displays and meters Equipment display OCR Camera capture for equipment without data ports OCR: CRAFT detection, CRNN recognition, edge devicesRelated project View related project: Reading equipment displays and meters Text from video CCTV meters, patent YOLO, SVTR OCR, edge devicesRelated project View related project: AI PDM master data generation AI PDM Items, BOM from requests LLM, VLM, human review RPA+OCR CompetitionRelated project View related project: LLM Compiler for equipment connectivity LLM Compiler R&D (2024 ~ 2026) Parsers generated from manuals, camera meter reading, interview setup, DAQ links, data modelling and SPC in one agent workflow LLM, SLM, OCR, source cross-checks, secure tunnel OCR+MLRelated project View related project: Automating orders and production records Intelligent RPA R&D Collect emailed orders and automate order-to-production records RPA, OCR, BERT document classificationRelated project View related project: Welding quality prediction Welding quality ML Current, voltage tuning Boosting, Bayesian optimization First AI projectRelated project View related project: Manufacturing data analytics validation Manufacturing data End-to-end analytics FFT features, boosting, AutoML, SHAPRelated project View related project: AI that connects manufacturing work ML engineer agent Data modelling, control charts set up for you Agent skills, control chart rulesRelated project View related project: AI that connects manufacturing work Conversational AI agent memory, tool use, screen drafts Human approval firstRelated project View related project: AI that connects manufacturing work workmanage Run work through workflows Harness, review nodesRelated project View related project: AI that connects manufacturing work AI report builder Reports by role Fixed queries, LLM notesRelated project View related project: Building manufacturing AI training datasets Data services Data voucher work, DNA specialist firmRelated project View related project: Building manufacturing AI training datasets Training datasets Weld quality, wind, robotic grasping NIA dataset projectsRelated project Smart DAQ development Equipment data collection gateway Modbus, MQTT, protocol conversion Q-Factory 3.0 SPA, cloud migration Cloud deployment base Q-Factory MES (since 2015) 1.0 deployment (since 2016), 2.0 standardization (since 2018), 3.0 rebuild (since 2024). Master data, BOM, inventory, production, quality and equipment. The base every AI reads and writes. VLM recognition Vision and OCR as one recognition model Detection, anomaly and keypoint heads on a base encoder Small VLM OCR together with a large VLM C-RADIO, PaddleOCR-VL, MinerU, Qwen Vision+OCR Document, drawing and CAD agent Read requests, manuals and drawings into items, BOM and inventory Generate 3D CAD from 2D drawings and text Build faster with in-house knowledge VLM, text to CAD, in-house knowledge search OCR+Vision Agent Workflow Set up by interview: cost allocation, MES setup, equipment links Workflow dashboard: person, team, alerts, reassignment, KPI ML skills: analysis steps, SPC and prediction Assist on every screen, reports checked by paragraph workflow engine, agent memory, TSFM, Tabular FM agent+ML Robot VLA Grown from robotic grasping data VLA Vision+AI data Unified agent platform VLM and LLM connect every area Domain AI does not disappear; absorbed as tools and skills People check the evidence and decide the next task Harness Every function as a tool or skill Agent-friendly backend APIs and databases agents read easily AI proxy, ZDR Calls that leave no data behind Bootloader A sandbox per person, with agent settings loaded in one sheet Subscription adapter Company proxy, own API key or own subscription (official login) Shop-floor context Smart glasses, CCTV, equipment data Agent engineering Guardrails, design patterns, eval, trace

The TO-BE column on the right is a research direction, not shipped work. After zooming in, drag the diagram to explore. Each project’s implementation status is listed below.

Related projects

Explore our manufacturing AI projects in chronological order.

14 projects

These AI-generated illustrations are based on company presentations and project materials. Equipment and screens are illustrative examples of how the systems work.

2018

Manufacturing AI in practice

2018 / Data analytics

Welding quality prediction

Research and development

We collected welding current and voltage data and analyzed how process conditions relate to weld quality.

AI-generated project illustration

Technologies Machine learning, process optimization

Save image

2019

Manufacturing AI in practice

2019 / Vision inspection

Pilot factory vision inspection

Deployed

We built a vision inspection system for a worker-friendly pilot smart factory. It captures parts and compares their detected positions with reference data.

AI-generated project illustration

Technologies Object detection, transfer learning

Save image

2020

Manufacturing AI in practice

2020 to 2022 / Vision inspection

Robotic vision for missing-part inspection

R&D and deployment

A robot moves to preset inspection positions to capture images. AI detects parts, compares them with reference data, and stores the results with supporting images.

AI-generated project illustration

Technologies Robot control, object detection, reference image comparison

Save image

2020 to 2021 / Vision inspection

PCB visual defect inspection

Pilot validation

Circuit boards are photographed under controlled lighting. The system compares them with reference images to locate visual defects.

AI-generated project illustration

Technologies Reference image comparison, anomaly detection

Save image

2020 to 2021 / Vision and text recognition

Food packaging and expiry date inspection

Pilot validation

The system checks package seals and contents for foreign material, reads printed expiry dates, and records the inspection results.

AI-generated project illustration

Technologies Image analysis, optical character recognition (OCR)

Save image

2020 to 2022 / Text recognition

Reading equipment displays and meters

Deployed

Cameras capture the displays of inspection equipment and meters that lack a communication interface. The system reads the numbers and saves them in the production management system.

AI-generated project illustration

Technologies Cameras, optical character recognition (OCR), equipment data collection

Save image

2020 to 2022 / Workflow automation

Automating orders and production records

Research and development

We researched automation that collects and classifies orders received by email and document, then connects them to production tasks and reporting.

AI-generated project illustration

Technologies Robotic process automation (RPA), OCR, document classification

Save image

2022

Manufacturing AI in practice

2022 to 2023 / Vision inspection

Poultry carcass quality grading

Deployed

The system captures poultry carcasses during production, assesses quality attributes, and stores grades and statistics. Image records extend quality checks beyond sample inspections.

AI-generated project illustration

Technologies Object detection, anomaly detection, body-region feature analysis

Save image

2022 to 2024 / Vision research

Exploring 3D reconstruction for weld inspection

Internal research

We explored methods for reconstructing a 3D object from images taken at multiple viewpoints. This work began with weld inspection and informed later research into visual understanding.

AI-generated project illustration

Technologies NeRF, Instant-NGP, DUSt3R

Save image

2022 / Training data

Building manufacturing AI training datasets

Dataset development

We collected and organized AI training data for applications including weld quality and robotic grasping. The work includes quality checks and preparation of training and evaluation datasets.

AI-generated project illustration

Technologies Data labeling, data quality verification, training dataset development

Save image

2023

Manufacturing AI in practice

2023 / Data analytics

Manufacturing data analytics validation

Pilot validation

We validated an analytics workflow that extracts features from equipment data and connects prediction models with explanations of their results.

AI-generated project illustration

Technologies Fast Fourier transform (FFT), boosting models, AutoML, SHAP

Save image

2024

Manufacturing AI in practice

2024 / Vision inspection

Extending missing-part inspection to bus frames

Deployed

We extended inspection to large frames by aligning camera positions with reference images and checking complex structures for missing parts.

AI-generated project illustration

Technologies Image segmentation, feature matching, reference image comparison

Save image

2024 to 2026 / Manufacturing AI research

LLM Compiler for equipment connectivity

In development

This research reads equipment manuals and communication documents to draft connection settings and data interpretations. Settings are checked against the source documents and tested with actual data collection.

AI-generated project illustration

Technologies Document recognition, language models, data acquisition integration

Save image

2025

AI PDM master data generation

2025 / Document AI

AI PDM master data generation

Field validation

The system reads items and specifications from development requests, drawings, and specification sheets. It drafts a bill of materials (BOM) and process information for a reviewer to check against the originals before saving.

AI-generated project illustration

Technologies Document and image understanding, structured output, human approval

Save image

TO-BE

The whole smart factory context, AI driven

We aim for a smart factory where AI leads manufacturing across product design and development, production planning, materials and inventory, quality, equipment, and cost management. AI understands factory-wide goals and conditions, decides and plans the work, and carries it out through business systems, equipment, and robots. Our TO-BE is a factory that continually improves by feeding execution results into its next decisions and plans.

How could AI fit your production process?

Let’s review your inspection needs, existing equipment, and repetitive tasks.

Discuss manufacturing AIBack to technology history