RLI am the co-founder and research lead at kaksha.ai, an AI-education startup where I focus on making AI accessible and beneficial for all. As a pragmatic, rational thinker, I provide strategic consulting to other AI startups, offering tailored solutions. I have strong skills in operationalizing the latest AI research, including cloud deployment of ML models, end-to-end lifecycle management of AI systems, and experience in creating RAG systems as well as fine-tuning LLMs. Further, I actively mentor students to get into AI, having mentored 8 so far.
Experience
AI Consultant (Part-time consulting)
IllumeetXP
Fine-tuned Open Source LLMs (Llama 2 and Mistral-7B). Curated a domain-specific dataset on Argilla and used PEFT to train LoRAs on the task reducing error rates by 35%. Deployed using AWS Sagemaker real-time and Modal serverless endpoints.
AI Engineer (Part-time consulting)
Botstacks.ai
Created the RAG system using a stack of LangChain and Pinecone, supporting interchangeable LLMs with adaptive prompts. Set up LLMOps using LangSmith which helped reduce bug turnaround time and saved ~$600/month by detecting bad actors.
Co-Founder
kaksha.ai
Leading the Research Team to operationalise EDM research through development of an internal library. Deployed state-of-the-art Deep Knowledge Tracing models (AUC 0.86) to over 10k students and 150+ teachers. Created and maintain data engineering pipelines to process 70mn+ datapoints. Scaled the company to MRR of ₹3 lakh (~$3k USD) without raising any investments.
Session Lead, Data Science Nanodegree
Udacity
Created and led supplementary sessions for 58 scholars. Crafted interventions for disengaged scholars leading a completion rate of 88%.
Education
University of California, San Diego
Bachelors of Science
Cognitive Science
Specialisation in Machine Learning. Minor in HCI.
Plaksha University
Tech Leaders Fellowship