AJ

AAKASHI JAISWAL

Applied AI Engineer

Bangalore, Karnataka, India

@aakashi

0 followers

About

Applied AI Engineer with experience in front-end development and founder’s office operations, plus hands-on work building RAG systems, AI coding agents, and multilingual voice AI applications. Comfortable working across product, engineering, and operations to ship practical AI-enabled software.

Experience

  • Founder's Office Intern
    AdroPardi
    Mar 2025 – Dec 2025

    Oversaw digital operations across website functionality, performance, user experience, and technical coordination. Collaborated with cross-functional teams spanning product, engineering, marketing, creative, and operations to translate business requirements into actionable deliverables. Supported platform improvements through content management, workflow coordination, and implementation of requested changes.

  • Front-end Engineer
    Teelure
    Sep 2024 – Dec 2024

    Maintained and improved an e-commerce website, implementing UI updates, product catalog changes, and usability improvements. Supported e-commerce operations involving inventory management, order processing, and logistics coordination. Collaborated on website updates and operational workflows to improve the online shopping experience.

Education

  • B.Tech
    Computer Science and Engineering (CSE) · 2022 – 2026

Skills

Tools / apps / platforms

<3 months <3 months Just getting started - under three months. This is also what shows when a level has not been set.
Git GitHub Ollama Claude Code

Projects

  • Bodh - Multilingual AI Voice Companion
    Python, FastAPI, SSE, Sarvam Saaras STT, sarvam-105b-conversations LLM, Bulbul TTS

    Multilingual voice AI application for English, Hindi, and code-mixed conversations with a streaming speech-to-LLM-to-speech pipeline and an offline evaluation suite.

  • NanoCode - AI Coding Agent Harness
    Python, OpenRouter

    Python-based AI coding agent with iterative LLM tool calling, file operations, shell execution, web tools, configurable model selection, permission-based execution, read-only Plan Mode, and context compaction.

  • DocuMind - Intelligent PDF RAG System
    React, Node.js, Qwen3 4B

    Full-stack RAG application for document search, conversational Q&A, and source-grounded answers. Improved BM25 Recall@6 from 42.3% to 84.6% on a synthetic dataset and added structured-output validation and source-quote verification.

🎯 Hobbies & interests

  • Reading
  • Running

Explore the Broxer community

🤖
Online · instant AI help
Broxer