![Banseog [ physical ai ]](/assets/brand/banseog-physical-ai-on-dark.svg?v=20261007){kind=small}
The company says some projects with CJ Logistics and Lotte have reached joint deployment. Once robots run inside a real work schedule, roles that design site operations together may become important.
CANDIDATE QUICK CHECK
30-second check before you apply
Check location, experience, education, work authorization, work mode and pay before opening individual jobs.
Current hiring data refreshes from official sources in the browser.
BANSEOG OUTLOOK
RLWRLD's shareholder list doubles as a list of the workplaces where its robots will learn
RLWRLD, founded in 2024 by Jung-Hee Ryu, builds AI models for robot hand work that can run on many kinds of robot. Its first funding, announced in April 2025, included LG Electronics, SK Telecom, KDDI, ANA Holdings, Mitsui Chemicals and Shimadzu. When it raised about $26 million more in February 2026, CJ Logistics, Lotte Ventures and Hyosung Ventures joined. The company says it trains its models inside the live workplaces of its strategic investors. In May 2026 it released its model and training code, and at its Seoul event in June, executives from Lotte Hotel, CJ Logistics and LG Innotek shared the stage. Banseog reads this as a company that weighs workplaces as heavily as money when it picks shareholders, and that keeps its model open while treating field records as its own asset.
Questions before joining
- Which shareholder sites does this role work with, and how often?
- Who owns the records collected at a workplace?
- What process applies when results learned at one shareholder's site are used at another's?
Logistics, hotels and factories differ in both objects and hand motions. As sites grow, deciding which records can be shared across industries and which stay apart may become important.
The company has spoken of a follow-up model and its Japan business. If the Tokyo base grows in role, receiving records in the procedures and language of Japanese sites may become important. These are conditional stages, not confirmed dates or hiring plans.
Capabilities gaining importance
Designing workplace data collectionNegotiating ownership and use of field dataRecording skilled workers' motion and forceCustomer-site deployment engineeringDesigning evaluation criteria for hand work
Talent markets to explore
Robot manipulation and imitation learning researchManufacturing and logistics automationCustomer-site deployment engineeringData collection hardware and sensorsTechnical sales to Japanese manufacturers
Current hiring comes from the existing verified data pipeline. This is a conditional talent and organization outlook, not a company hiring plan, personal-fit assessment or hiring-probability forecast. · 2026-11-01
FUTURE TALENT SHIFT
How responsibilities may expand
Follow the responsibility that runs from model performance in the lab to a day's work at a shareholder's site. Check individual postings for actual duties and qualifications.
Improving public benchmark scores→Judging model design with the limits of workplace records in mind
Cleaning training data and managing quality→Managing records from different industries under one format and access standard
Building and supporting demo setups→Deciding with the shareholder which part of the process the robot takes
Finding joint projects with investors→Also negotiating which records come from which site and when
Reviewing general contracts and IP→Designing ownership of field records and their use at other shareholders' sites
Porting the model to one robot→Absorbing differences so the same task works on robots from other makers
CAREER LENS
Experience turning people's hand skills into records a machine can learn from
Beyond building good models, look for times you went into a workplace and recorded how people actually work. Adjacent experience suggests connections, not eligibility or a personal-fit verdict.
Experience to show in your career history
✓
Working on site at a factory or logistics center to analyze tasks
✓
Measuring and recording skilled workers' motion or force
✓
Training a robot to follow human demonstrations
✓
Collecting data within a customer site's safety and security rules
✓
Applying one control model across different robots
✓
Running technical projects with Japanese or other overseas manufacturers
Where adjacent experience can lead
Breaking work into motions and timing it can connect to record design.
Field experience handling irregular objects is the connection.
This can lead to adding force and touch to visual understanding.
Experience measuring finger contact and force.
Solving problems at customer sites can connect to deployment roles.
Knowing the procedures and exceptions of repetitive work can help in site design.
Questions before joining
- Which shareholder sites does this role work with, and how often?
- Who owns the records collected at a workplace?
- What process applies when results learned at one shareholder's site are used at another's?
- Who decides the line between the open model and assets kept private?
- Is performance judged on research results or on deployment results?
PLACE × TALENT INTELLIGENCE
Work connected across locations
These interpretations use publicly reported locations and business activity. They do not describe regional hiring plans or current staffing.
Public job information lists an office in Gangnam, Seoul. This reads as where records from shareholder sites come together, with no estimate of current team structure.
Robot AI researchData preparationBusiness development
Information on the Japan entity gives an address in Marunouchi, Tokyo. This reads as the door through which Japanese workplaces open; check postings for actual placement.
Japan business developmentTechnical proposalsSite coordination
CJ Logistics has said it plans to co-develop a logistics model and apply it to its centers. This reads as the stage for learning irregular objects, without presuming the scale of use.
Logistics task analysisRobot deploymentOperating records
At the June 2026 event a Lotte Hotel executive described ideas such as room item delivery and washing tableware. No room cleaning demo was shown, so this reads as an idea stage.
Service task analysisIrregular objectsWorking alongside people
Shareholder investment → workplace access → motion and force records → model training → site deployment → the next shareholder joins