Selected to contribute to Anthropic's AI safety initiatives.Signed an NDA 📄🖋️ so 🤫

▪𝟭⁃𝟬𝟬𝟳 𝗔𝗜 𝗔𝗴𝗲𝗻𝘁:Working on AI Agents equipped with Retrieval Augmented Generation & Unicode symbolectic function calling using the latest reasoning models like o1.▪𝗠𝗮𝗿𝗸𝗲𝘁𝗠𝗮𝘃𝗲𝗻 ⁃ 𝗚𝗲𝗻𝗲𝗿𝗮𝘁𝗶𝘃𝗲 𝗧𝗿𝗮𝗱𝗶𝗻𝗴 𝗠𝗼𝗱𝗲𝗹𝘀:Developed trading models using 𝘓𝘚𝘛𝘔𝘴, 𝘛𝘦𝘯𝘴𝘰𝘳𝘍𝘭𝘰𝘸 𝘈𝘨𝘦𝘯𝘵𝘴, 𝘚𝘵𝘢𝘣𝘭𝘦 𝘉𝘢𝘴𝘦𝘭𝘪𝘯𝘦𝘴3, 𝘛𝘳𝘢𝘥𝘪𝘯𝘨 𝘢𝘯𝘥 𝘝𝘪𝘴𝘪𝘰𝘯 𝘛𝘳𝘢𝘯𝘴𝘧𝘰𝘳𝘮𝘦𝘳𝘴 (𝘓𝘓𝘈𝘝𝘈, 𝘎𝘗𝘛-4), and 𝘵𝘦𝘹𝘵 𝘵𝘳𝘢𝘯𝘴𝘧𝘰𝘳𝘮𝘦𝘳𝘴 (𝘓𝘓𝘈𝘔𝘈, 𝘔𝘪𝘴𝘵𝘳𝘢𝘭 𝘦𝘵𝘤).▪𝗖𝗼𝗻𝘁𝗲𝘅𝘁𝘂𝗮𝗹 𝗧𝗿𝗮𝗻𝘀𝗹𝗮𝘁𝗼𝗿 𝗖𝗵𝗮𝘁𝗿𝗼𝗼𝗺:Used 𝘎𝘗𝘛-4 API for contextual language translation for a multilingual chatroom.▪𝗣𝗲𝘁 𝗦𝗮𝗳𝗲𝘁𝘆 𝗖𝗮𝗺𝗲𝗿𝗮 𝗠𝗟 𝗠𝗼𝗱𝗲𝗹𝗶𝗻𝗴 ❪𝗜𝗻𝘁𝗲𝗿𝗻𝘀𝗵𝗶𝗽 𝗣𝗿𝗼𝗷𝗲𝗰𝘁❫:Implemented 𝘠𝘖𝘓𝘖, 𝘚𝘪𝘢𝘮𝘦𝘴𝘦 𝘕𝘦𝘵𝘸𝘰𝘳𝘬𝘴, and 𝘉𝘓𝘐𝘗 models for pet identification and action tracking in a pet safety camera project.

▪𝗔𝗜 𝗖𝗼𝗻𝘁𝗲𝗻𝘁 𝗗𝗲𝘁𝗲𝗰𝘁𝗶𝗼𝗻 𝗠𝗼𝗱𝗲𝗹 𝗗𝗲𝘃𝗲𝗹𝗼𝗽𝗺𝗲𝗻𝘁:Spearheaded the creation of AI content detection models, contributing to advancements in ethical AI practices. This work gained recognition, including a feature in a news interview on 𝘔𝘚𝘗 𝘔𝘦𝘥𝘪𝘢 𝘕𝘦𝘵𝘸𝘰𝘳𝘬'𝘴 𝘈𝘐 𝘙𝘰𝘶𝘯𝘥 𝘜𝘱.▪𝗔𝗰𝘁𝗶𝘃𝗲 𝗣𝗮𝗿𝘁𝗶𝗰𝗶𝗽𝗮𝘁𝗶𝗼𝗻 𝗶𝗻 𝗣𝗿𝗲𝘀𝘁𝗶𝗴𝗶𝗼𝘂𝘀 𝗔𝗜 𝗖𝗵𝗮𝗹𝗹𝗲𝗻𝗴𝗲𝘀 𝗮𝗻𝗱 𝗙𝗼𝗿𝘂𝗺𝘀:Engaged 𝘌𝘵𝘩𝘪𝘤𝘢𝘭𝘨𝘰 in significant AI initiatives, including 𝘖𝘱𝘦𝘯𝘈𝘐'𝘴 𝘗𝘳𝘦𝘱𝘢𝘳𝘦𝘥𝘯𝘦𝘴𝘴 𝘊𝘩𝘢𝘭𝘭𝘦𝘯𝘨𝘦, 𝘔𝘐𝘛𝘹 𝘚𝘰𝘭𝘷𝘦, 𝘈𝘐 𝘈𝘭𝘪𝘨𝘯𝘮𝘦𝘯𝘵 𝘈𝘸𝘢𝘳𝘥𝘴 and 𝘙𝘐𝘚𝘌, demonstrating commitment to AI innovation and ethical solutions.▪𝗩𝘂𝗹𝗻𝗲𝗿𝗮𝗯𝗶𝗹𝗶𝘁𝘆 𝗗𝗶𝘀𝗰𝗹𝗼𝘀𝘂𝗿𝗲:Reported a major vulnerability for 𝘌𝘭𝘦𝘷𝘦𝘯𝘓𝘢𝘣𝘴 (free generation token exploit) and a prompt leakage exploit (non-standard Unicode) for 𝘒𝘩𝘢𝘯 𝘈𝘤𝘢𝘥𝘦𝘮𝘺’𝘴 𝘒𝘩𝘢𝘯𝘮𝘪𝘨𝘰 & other Custom GPTs.

▪𝗣𝗿𝗼𝗺𝗽𝘁 𝗗𝗲𝘃𝗲𝗹𝗼𝗽𝗺𝗲𝗻𝘁 𝗳𝗼𝗿 𝗖𝗼𝗻𝘁𝗲𝗻𝘁 𝗪𝗿𝗶𝘁𝗶𝗻𝗴:Created natural-sounding, high-quality content prompts tailored to client needs, enhancing the effectiveness and engagement of their content strategies.▪𝗩𝗼𝗹𝘂𝗻𝘁𝗮𝗿𝘆 𝗖𝗼𝗻𝘁𝗿𝗶𝗯𝘂𝘁𝗶𝗼𝗻𝘀 𝘁𝗼 𝗔𝗜 𝗦𝘆𝘀𝘁𝗲𝗺 𝗜𝗻𝘁𝗲𝗴𝗿𝗶𝘁𝘆:Volunteered for 𝘔𝘪𝘥𝘫𝘰𝘶𝘳𝘯𝘦𝘺, focusing on identifying and reporting adversarial prompts & jailbreaks.