Alchemyst AI, a generative AI startup based in Bengaluru, has successfully raised $300K in its pre-seed round from prominent investors such as Inflection Point Ventures, 100 Unicorns (formerly 9U), and EarlySeed Ventures. The company’s mission is to create AI employees that can assist humans in solving high-level intelligent problems in enterprises. In an exclusive interview, founder Uttaran Nayak shared that Alchemyst AI has developed its own model, AlchemystC1, which combines Phi-3 vision and an Indic LLM from AI4Bharat. The company plans to expand into 28 languages in the next few months to cater to Indian customers. To address concerns about data exposure, Alchemyst AI does not use popular models like OpenAI’s GPT-4o or Google Gemini. Instead, they have fine-tuned their model to say “I don’t know” rather than producing gibberish. The company’s chief technology officer, Anuran Roy, explained that they have developed a dedicated mesh architecture for storing and auditing data over time. Customers can also define scopes to restrict LLMs from accessing certain data. Alchemyst AI guarantees that no data prompts leave the company and is compliant with strict standards such as GDPR, HIPAA, and SOC. The startup currently uses multiple cloud service providers for LLM inference and is in talks with E2E Networks to acquire more GPUs. Founded by school friends Nayak and Roy, Alchemyst AI officially launched in November 2023 and currently has a team of six technical experts. Nayak was inspired to start the company while working with a B2B SaaS company during his third year of engineering.
