Game AI: deep reinforcement learning, dynamic difficulty adjustment, and draft recommendation in multiplayer online battle arenas. Advisors: Farzin Yaghmaee, Morteza Dorrigiv.
Education
Experience
Run the studio: direction, team, and the commercial side of shipping mobile games to the Iranian market.
Technical lead across the studio's catalogue — Unity, C#, backend services, multi-store publishing and in-app purchase across CafeBazaar, Myket and others.
Computer Fundamentals, Advanced Programming, Multimedia, Enterprise Application Integration, Computer Workshop.
Enterprise Application Integration, Network Laboratory, Operating Systems Laboratory.
Artificial Intelligence, Network Laboratory, Graphics Laboratory.
Artificial Intelligence, Algorithm Design, Data Structures, Computer Fundamentals and Programming, IT Engineering. Computer and network laboratory assistant.
Selected publications
Twelve in total — the full list is on Google Scholar.
45 citations. Training an agent to play through deep reinforcement learning, evaluated in a released commercial game. PPO gave the strongest result.
Games shipped
Eight titles on Google Play as SinSin Studios, and on CafeBazaar and Myket for the Iranian market. Unity and C# throughout, with the backend services, realtime multiplayer and in-app purchase behind them — and, on the studio side, the team and the release schedule. More at sinsinstudios.ir.
500,000+ downloads. A crossword game with a puzzle catalogue that keeps growing, and a sequel built on what the first one taught us about how people actually solve.
500,000+ downloads. Online play with a shared catalogue — the same puzzles, the same evening, several people.
Ludo against friends or strangers in real time, with matchmaking and a leaderboard.
Hokm against computer opponents, playable with no connection at all — which is most of the engineering, and all of the reason it gets played on the metro.
Published on CafeBazaar.
Card, tile-matching and puzzle titles. Dragon Cave doubled as the test environment for the deep reinforcement learning agent in the 2019 paper — the research and the catalogue are not separate things.