Flowline is built by Dinan Elsyad and Berke Altiparmak for the overlooked translation work inside civil infrastructure design: turning CAD, GIS, plan sheets, standards, and engineering judgment into reviewable schedules, QA checks, and tradeoff outputs.
Flowline
Decision support for design teams
Flowline needs people who can respect expert workflows while making complicated technical material usable. That is the founding lane.
Turns messy expert data into tools people can actually use.
Dinan studied Statistics at Harvard and has worked deeply with survey-scale data: published research, an R package built for FIA and federal workflows, and tooling that makes complex datasets usable for expert users.
Builds AI systems that connect reasoning, data, and product speed.
Berke studies Data Science at Harvard, brings hands-on AI and LLM depth, and has quant experience that sharpens how Flowline thinks about evidence, uncertainty, and high-stakes decision support.
Flowline keeps sources, confidence, flags, and assumptions visible instead of pretending the machine has final authority.
The team is learning from transportation practitioners working around Bentley-centered DOT deliverables, not generic CAD demos.
The wedge is practical: read project artifacts, structure the useful information, and help engineers move faster with evidence.
Flowline is focused on the connective tissue around design: extracting the information already present across project files, plan sheets, GIS layers, reports, and specifications, then turning it into outputs an engineer can review and trust.
Current wedge