I build AI that decides who and what to trust online
I study how bots and AI agents behave, design the ML and agent systems that catch abuse, and make the calls on what to ship. The real work is deciding what to allow, challenge, block or charge, and who pays when it's wrong.
What I work on
-
01
Fraud and risk
Building the systems that decide whether a click, payment, account or listing is fraudulent, including the AI agents that now review cases and enforce policy.
-
02
Bot and agent traffic
Telling people, good bots, verified AI agents and abusers apart, request by request, and deciding what each one may do.
-
03
Agent security
Keeping the AI agents a company runs from being hijacked, misused or trusted with too much, and deciding what each one may access.
About me
ML engineer turned AI product leader
I am Syed Tanveer Jishan, a Bangladeshi-Canadian based in Vancouver and Chief Innovation Officer at MonetizeMore. I have spent 10+ years in machine learning and AI products, nine of them building systems that tell real people from bots and fraud in digital advertising. I now lead agentic AI products from idea to production, and I hold an MSc in Computing Science from SFU, with published ML research.
I write on LinkedIn about the patterns bots and agents leave in the data, how to design the systems that catch abuse, and the calls that decide what ships. For advisory work or speaking, email me or message me on LinkedIn.
On record
- 8.7% to 1.1% Ad-network revenue deductions, after Traffic Cop moved from heuristics to ML. It now protects more than $2M a year in ad spend.
- 86.3% Compliance agent accuracy on a 250-site benchmark covering 102 policy checks. It launched only after passing.
- 99.9% Accuracy bar on a payment data platform handling $4 to 6M a month across 80+ networks.
- −65% Technology operating expense, cut while I rebuilt engineering and product leadership.
Two Google Publishing Innovation Awards (2020, 2022), for Traffic Cop's impact on ad fraud prevention.