Niraj Patel

Every enterprise running GenAI in production hits the same wall: the model gets it wrong, and there's no infrastructure to catch and correct the mistake before it reaches the customer. Existing guardrails vendors flag, score, or block. None of them fix. DeepRails is the only platform in the world that does. We evaluate AI outputs against a suite of guardrails (correctness, safety, completeness, instruction adherence, custom domain metrics), automatically remediate failures, and verify the corrected version before it ships. Our HallucinationSafe™ certification is the SSL of AI: a verifiable trust mark for AI outputs in production.

Member of Anthropic's invite-only Model Safety Program. I evaluate the latest, unreleased Claude models for adversarial robustness, with an emphasis on high-risk domains.

• Reported to Trilogy's group CEO & CTO. Responsible for building 2 edtech software businesses from the ground up, namely GT School & LearnWith.AI, both of which have products that were featured in Forbes and US national media. • Oversaw a $10MM budget – hired & managed 60 direct reports across 6 teams (AI Prompt Engineering, Product, AI-first Marketing, AI-first Operations, Domain Experts/SMEs, & Customer Success). • Hands-on leader - defined the strategy & product roadmap, daily work tasks (JIRA), and ran daily stand-ups for all 6 teams. Ensured continuous gen AI innovation & the shipping of 1 major product feature each week. • Developed 3+ AI-powered software products (B2C) and built a set of AI Education APIs (B2B). Examples incl. TeachTap (iOS), Magic Academy (iOS, Android, PC), and TeachTales (iOS). All contain patented applications of generative AI both in terms of generative capabilities and learning innovation. • Performed continuous generative AI focused R&D for product discovery purposes for text, image, video, and audio. Discovered several key breakthroughs that resulted in our subsidiary rebranding entirely; published research company-wide to foster responsible AI adoption. • Continuously analyzed user behavior in Mixpanel to adjust product & marketing strategy – reduced CAC by 90%, CMQL by 85%, and increased conversions by >5,000% -- driven primarily through delivering innovative gen AI features. • Successfully launched a generative AI-first marketing strategy on TikTok, Meta, Google, and Snapchat, creating an audience of 250k in 2 mos. • Presented weekly data-backed update reports to CTO & EVP Technical Product Management to gain alignment & share progress on shipping new gen AI product features, and progress on the adoption/use of AI internally.

• Created and implemented machine learning algorithms that enhanced risk assessment and pricing strategies for multinational insurers, resulting in a 20% improvement in predictive accuracy and supporting clients’ sustainability initiatives. • Spearheaded the adoption of AI and blockchain technologies within actuarial processes, enabling real-time data analysis and sustainable reporting for financial institutions, which contributed to earning three industry awards for innovation and digital transformation. • Built several pricing models & algorithms in Python & JavaScript for Lloyd’s of London & other multi-national insurers. • Conducted 5 actuarial-focused financial audits of multinational insurance companies. • Pioneered adoption of AI/ML & Blockchain in financial services - developed several full-stack prototypes and presented them to large audiences. Resulted in 3 industry awards for innovation and digital transformation.

AI/ML, Stochastic Calc, Math in CS, Adv. Statistical Methods

Part qualified actuary