How Shigeo Rebuilt a Data Career—and Walked Into Cohere

Shigeo’s first taste of “abroad” was not a job offer. During a Kansai University informatics master’s he presented a poster at UIST 2017 in Québec City—his first time answering questions in English on the spot, with exam English as his main preparation and a professor nearby for backup. The surprise was that he could manage. SIGGRAPH student volunteering in Vancouver in 2018 deepened the feeling that overseas work was reachable, especially after a Canadian-based Chinese volunteer lead looked after him and later introduced friends—and, eventually, a recruitment contact—even though that lead’s own visa later ended and he returned to China before Shigeo arrived.

After the master’s he joined Supership, a KDDI-group adtech firm, as a new graduate. For about three and a half years he built the ScaleOut DSP dashboard in Rails and React—rewriting jQuery UI and upgrading Rails. He had chosen the company partly for its data assets; when an internal data-engineering seat opened, he moved under a “T-shaped talent” policy he and the company both wanted. Roughly a year later he was seconded—not transferred—to UltraImpression, a joint venture with TV Asahi and Dentsu, as data-team leader among secondees and contractors. He had wanted management experience; leading people more senior than him on contract also left a technical gap he wanted to close.

Vancouver won over Australia and other working-holiday destinations because of those SIGGRAPH ties and a sense of tech opportunity. A general agency pushed working holiday only and seemed not to understand engineering work; Frog helped him compare working holiday with a Co-op college track. He chose CICCC’s data-science Co-op path—nine months of classes, nine months Co-op—to rebuild foundations across data engineering, analysis, and science, finally enter Kaggle competitions, and practice English presentations. The cohort surprised him: many classmates were there mainly for visa runway, and an instructor vanished after day two. Even the career advisor, he recalls, did not know what a data engineer was.

He started applying about two months after an August 2024 arrival, deep in a soft market. Early volume—on the order of a hundred tries—brought almost no interviews. Across the search he would reach about two hundred applications, supporting living costs with Japan contracting work while he persisted. Permanent residency stayed conditional: he was open to returning to Japan if Canada did not fit, and he noticed food and employment stability still pulled him toward home.

The role that closed the loop was Senior Data Quality Specialist at Cohere, working on global product quality for an LLM company. It sits at the intersection of the data craft he had expanded in Japan and the international stage he first tasted at academic conferences—earned, in a recession year, by volume, side income, and refusing to treat one quiet autumn as the end of the story.

He is careful about how he weighs Canada against Japan now. Layoff talk makes employment feel less stable than Japan’s seller’s market; rice and home cooking still win on taste. Permanent residency remains a conditional idea—keep it if the fit is real, return if it is not—while the Cohere title shows the data axis he spent years widening can still open a global product door.