krishm.dev

Krish Makhijani

rss feed

Currently coding
with

Hey 👋 Krish here! I've been coding in Linux primarily for over 5 years. Currently learning Langchain and GoLang

github - https://github.com/krishmakhijanitwitter - https://x.com/heykrishhlinkedin - https://linkedin.com/in/krishmakhijanimedium - https://medium.com/@krishmakhijanimail - mailto:me@krishm.devdiscord - https://discordapp.com/users/771616688407445515

Languages: C, C++, Java, Python, Javascript, Typescript, Bash

Frontend: React, Next.js, Tailwindcss, Framer motion, Shadcn, Zustand

Backend: Node, Bun, PostgreSQL, Supabase, ChromaDB, Pinecone, MongoDB, Prisma, Drizzle, Redis, Apache Kafka, AWS, DigitalOcean, Hono, Express, Flask, FastAPI

DevOps: Docker, Kubernetes, Jenkins, GitHub Actions, Terraform, Ansible, Grafana, Prometheus

WEB3: Solana, Etherium, Smart Contracts, Solana Wallet Provider, Solana Mobile Wallet Provider

Generative AI: Retrieval-Augmented Generation (RAG), Langchain, Langraph, Cohere, LlamaIndex, , Vercel AI SDK

Machine Learning: PyTorch, OpenCV, Yolo

Tools: Git, BitBucket

Discord Status
Projects
Blogs
Experience

Gaido AIRemote

March 2025 - May 2025

Software Engineer Intern

  • Developed a logic that generates and stores OpenAI Embeddings of the essential insurances and stores them into the database.
  • Developed the HNSW indexing mechanism to enhance the efficiency and speed of context retrieval from the database.
  • Created a Vector Similarity Search Engine utilizing the Cosine Similarity Algorithm to retrieve relevant documents from the database and rank the top two documents using the Cohere Ranking Function.
  • Streamlined the insurance selection process by developing the core logic of the onboarding agent, which captures User Details, updates the user UserProfile State, and stores the information in the database. Which helps in the quick selection of the best health insurance based upon the User Profile State

Samsung IndiaBengaluru, Karnataka, India

July 2024 - March 2025

Research and Development Intern

  • Designed a Smart Battery Prediction System to improve android smartphone user experience by predicting battery discharge rates accurately.
  • The system takes input from data provided by android smartphone sensors like screen on/off, app usage, and network connectivity and makes time-series predictions based on advanced LSTM and xLSTM models
  • Resolved the problem of data continuity through a RandomForestClassifier to support quality predictions.
  • The project is a blend of machine learning and mobile analytics and provides end users with practical suggestions for battery efficiency and improved device optimisation.