Summary

Heya! I’m a back-end engineer with an experience in working with LLMs for AI start-ups and in fin-tech start-ups. I have built custom AI chatbots, fine-tuning models and developing strategies for trading algorithms.

Work experience:

• 🔗 RedTree.ai:
  • I served as the lead engineer in the development of a platform which enabled users to create Personalised AI chatbots for their websites, leveraging their custom data to enhance customer engagement. Users can upload thousands of pages of documents, empowering the AI chatbot to effectively address queries related to the uploaded documents.
  • I was the core engineer responsible for developing the majority of features for our other product, Mr. Owl, an AI-powered guide serving as both a tutor for learners and an assistant for teachers.
  • Configured and set up an entire cloud infrastructure & services(Database, Authentication, storage buckets, edge functions) utilising Supabase.
  • Implemented PostgreSQL database for complex data relationships , Supabase Javascript Library for API infrastructure, Vercel AI SDK with Next.js and OpenAI to create a ChatGPT-like AI-powered streaming chat bot.
  • Developed and integrated an AI tutor Chrome Extension designed to assist students in asking questions and clarifying doubts while coding on the Replit platform.
  • Employed smooth & scalable UI development using ShadCN UI, built-on TailwindCSS and Radix UI.
  • Researched and built Voice-to-Voice and Text-to-Voice features(Accuracy >99%) on the platform by leveraging OpenAI Models (GPT-4 & Whisper) and the MediaStream Recording API from Mozilla. You can checkout the related blog 🔗 here.
  • Migrated the entire backend for the extension to Fast API in order to comply with Google Developer Policy.
• ModelsLab
  • Streamlined user experience by developing a one-click Docker deployment system.
  • Managed GPU allocation on our local server for high-processing LLM containers.
  • Implemented a Network File Server for efficient Docker volume management, enhancing data storage and retrieval capabilities in the backend infrastructure.
  • I Was Responsible for Shifting our infrastructure off the cloud to our own GPU infrastructure to save costs for training LLM models.
  • Implemented a Python-Docker SDK interface facilitating two-way communication between the backend web server and a local server where Docker images were running.
  • Built the back-end infrastructure with Django, Python with Typing, SISH, bash, Linux, SQL Alchemy, Alembic, Postgres, Docker, Stripe, Microsoft Azure
• 🔗 PolyByte
  • Developed high-performance trading systems by adding real-time data feeds, trading algorithms, order management, backtesting, and integration with data providers through APIs.
  • Improved stocks back-testing features and database performance.
  • Crafted an intuitive frontend interface, empowering users to seamlessly monitor live market trends, analyse trading strategies, and execute trades with precision.