# Randunu Dimeshan, AI Product Manager > Randunu Dimeshan is a Product Manager based in Colombo, Sri Lanka, with 8+ years of experience taking AI, cloud and platform products from launch to scale for clients in Europe, the Gulf (MENA) and North America. Randunu is open to consulting and product advisory work. Contact: randunu1483@gmail.com. ## Summary - Role: Product Manager (Manager, Product Management) at Type B Digital (Canada, remote), since October 2024. - Location: Colombo, Sri Lanka. Works remotely across time zones. - Focus areas: AI and generative AI products (LLMs), PropTech, FinTech and private equity platforms, 0-to-1 MVPs, platform modernization, delivery process design. - Open to: consulting and product advisory engagements. - Certifications: Certified Scrum Product Owner (CSPO), Certified ScrumMaster (CSM), AWS Cloud Practitioner Essentials, Google Project Management Certificate, Product School micro-certifications. - Education: BSc (Hons) Business Information Systems, Cardiff Metropolitan University (UK); Higher Diploma in Software Engineering, ICBT College (Sri Lanka). - Also: PADI Divemaster and Master Scuba Diver; built the SL Turtle ID Project, a sea turtle photo-identification platform. ## Highlights - Led product pods for an AI real estate platform that launched to 100+ paying brokerages and supported a $7M seed round (3x the pre-seed). - Built a real estate equity platform from 0 to 1, delivering the MVP in 16 weeks and supporting SEC compliance approvals. - Modernized a 9-year-old private equity platform with 300+ features; cut system response times by 70%. - Helped Type B Digital expand with new clients and 50 hires across Europe, the Gulf and South America. - Supervised and coached 20+ designers, engineers and product managers. - At Code94 Labs: delivered 25+ WordPress projects (LKR 8.5M revenue), launched 8 web and mobile product suites, reached 70% client retention, helped win 50+ new clients, rolled out Scrum to a 20+ person engineering team. - At Polhena Diving Center: 150% revenue growth in year one and about 30% a year after; 90% of revenue from digital channels. ## Case study 1: Making an AI creative engine reliable enough to launch - Product: A full-service AI marketing assistant for real estate agents and agencies (client name withheld). - Role: Product Manager. Team: 2 engineers, 1 designer, working across a 12.5-hour time difference. The product: Each agent gets a personal assistant tuned to their brand and tone of voice. It creates social posts, graphics and email newsletters, posts on a schedule and helps find leads. On sign-up it pulls in all of the agent's MLS listings, keeps them updated, and builds a posting schedule automatically. The problem: The product and waitlist were strong, with hundreds of brokerages waiting, but the core creative engine was not deterministic: the same input produced slightly different output each time. There were close to 1,000 unstructured work items, tight deadlines, frequent demos, and manual QA that relied on people visually comparing output with reference designs. What Randunu did: - Rebuilt the engine around a templating system that gives the same result every time while staying customizable per brand. - Created an intake-to-delivery process so every request had a clear path from idea to demo. - Moved the team to small work cycles, shorter than sprints, each ending in a usable, demo-ready piece. - Built an internal automated QA tool that checks generated designs against reference designs. - Set core overlap hours and an async routine across a 12.5-hour time difference. Results: The core engine shipped. The client raised a seed round 3x its pre-seed ($7M) and onboarded 100+ paying brokerages, including some of the top brokerages in their country. Lesson learned: You don't always need to run an extremely tight ship. The startup moved at lightning speed; the team built solid processes where they mattered (intake, QA, demo-ready cycles) but cut red tape so people could move fast, especially before demos. Some chaos is fine when teammates are experienced. ## Case study 2: Modernizing a 9-year-old investment platform with 300+ features - Product: A finance and private equity platform serving Europe and MENA, expanding to the US (client name withheld). - Role: Product Manager. Team: 6-8 front-end engineers, 2 back-end engineers, 2-3 designers. The problem: Over nine years, client-specific features had been built straight into the core product instead of configured per client, leaving 300+ features to modernize on a short timeline. Team turnover meant missing context, there was no API or technical documentation, and the app was mostly server-side rendered with few REST APIs. What Randunu did: - Used AI to reverse-engineer how the backend and its financial calculations behaved. - Documented 300+ features end to end for the first time. - Built AI skills for QA runs, PRD writing and design checks. - Automated a Jira backlog of close to 10,000 tickets so the team always knew where the project stood. Trade-offs: The team chose not to do a complete rewrite. The foundations were never built correctly, and fixing them properly required changing the underlying infrastructure, which the timeline did not allow. Deep backend rework was deferred; within those limits, targeted backend fixes cut system response times by 70%. What Randunu would do differently: fix the backend first and standardize calculations into one library; bring AI in on day one with a shared "project brain", skills and work plans; and rebuild less, shipping a smaller, 100% accurate and faster core to win new clients instead of rebuilding every feature. ## Experience - Product Manager (Manager, Product Management), Type B Digital, Canada (remote), Oct 2024 to present. - Lead Project Manager, Code94 Labs, Sri Lanka (remote), May 2023 to Sep 2024. - Operations Manager, Polhena Diving Center, Sri Lanka (hybrid), Jan 2017 to Apr 2023. ## Skills - Product: product strategy, roadmapping, discovery, go-to-market, requirements, stakeholder management. - Delivery: Agile, Scrum, Kanban, sprint planning, backlog management, process optimization, team leadership, coaching. - Technical: AI and machine learning (LLMs, generative AI), AWS, GCP, UI/UX, platform infrastructure, API integration. - Business: client acquisition, hiring, revenue growth, P&L awareness, negotiation, vendor management. - Tools: Jira, Linear, ClickUp, Figma, FigJam, Notion, Coda, GitHub, Vercel, v0, Bolt, Lovable, Google Analytics, Google Tag Manager, Hotjar. ## About Before tech, Randunu helped their uncle build a scuba diving resort on Sri Lanka's south coast, with conservation at its heart. There Randunu built the SL Turtle ID Project, a crowdsourced turtle tracker that uses pattern recognition to identify individual sea turtles; it has catalogued 100+ turtles and was featured by Mongabay. Randunu's dream job is building product at a blue carbon or marine conservation startup. Outside work: reading and gaming. ## Links - Website: https://randunudimeshan.com/ - CV (PDF): https://randunudimeshan.com/Randunu_Dimeshan_CV.pdf - LinkedIn: https://www.linkedin.com/in/randunu/ - Medium: https://medium.com/@randunudimeshan - Mongabay feature on SL Turtle ID: https://news.mongabay.com/2021/08/in-sri-lanka-biologists-and-divers-build-a-facebook-for-sea-turtles/ - Email: randunu1483@gmail.com