Mihir Bhoi

Hi I'm Mihir,

I build applied AI systems end to end — from a TensorFlow disease-prediction model to Archis, an AI-native mobile browser with an on-device/cloud LLM architecture, tool-use orchestration, and safety guardrails.

About Me

I'm a software engineer moving deliberately into ML Engineering and Forward Deployed Engineering. My hands-on ML work started with a TensorFlow neural network that predicts disease probability from symptoms, and has grown into Archis, a mobile AI browser I designed and built solo — a tiered on-device/cloud LLM architecture with an agentic orchestrator that plans, calls tools, and completes tasks inside real web pages under explicit safety guardrails.

Before that, I spent two years in technical support, diagnosing issues directly with customers and tracing them to root cause — the same instinct for embedding with users and shipping working systems that Forward Deployed Engineering demands. I like sitting at the intersection of applied ML and production mobile engineering: making models and agents actually work inside a real product, not just in a notebook.

bhoi.mihir7@gmail.com
Gujarat, India
2+
Years Building Software
3
LLM Providers Integrated
5+
Agent Task Pipelines Shipped

Explore My Expertise

Filter skills and projects by role

Skills & Technologies

AI / ML

Machine Learning
Neural Networks
TensorFlow
LLM Integration (Claude, Gemini, OpenAI)
On-Device LLMs (llama.cpp)
Apple Foundation Models
Agentic Orchestration & Tool Use
Prompt Engineering
AI Safety Guardrails

Programming Languages

Python
TypeScript
JavaScript
Java
Kotlin
Swift
SQL

Mobile & Systems Engineering

React Native
Expo
Android Studio
Native Modules (Swift / Kotlin)
SQLite
Zustand
Android Jetpack
SwiftUI

Cloud, APIs & Data

Firebase
REST APIs
Node.js
MySQL
MongoDB

Tools

Git
GitHub
Xcode
Visual Studio Code
IntelliJ IDEA
Postman

Professional Skills

Customer-Facing Troubleshooting
Root-Cause Analysis
Communication
Cross-Functional Collaboration
Documentation
Agile / Scrum

Featured Projects

Applied ML and agentic AI systems, backed by hands-on mobile engineering.

Archis Browser logo
Archis Browser
Flagship Project

Personal project · Solo-built · iOS & Android

An AI-native mobile browser with an integrated "AI Mode" that reads, summarizes, and operates web pages on the user's behalf — designed and built solo, including the native app shell and the full AI subsystem.
  • Three-tier AI architecture chosen per task by cost and privacy: deterministic on-device extraction, an on-device LLM (Apple Foundation Models or a local GGUF model via llama.cpp), and connected cloud models (Claude, Gemini, OpenAI) with a self-healing model catalog.
  • An agentic orchestrator that plans, calls browser tools through a WebView bridge, asks clarifying questions, replans, and only reports a verified task completion.
  • Safety enforced in code rather than the prompt: step/cost/time budgets, domain and financial-action blocklists, an approval gate before purchases or submitting personal data, and an audit log of every run.
  • Purpose-built agent pipelines for shopping price comparison, coupon and wallet-offer discovery, travel search, page research, and form autofill — each degrading gracefully instead of failing the task outright.
  • Supporting product surfaces built end to end: an encrypted password vault, cross-device sync, and a native ad/content-blocking engine with compiled, platform-validated blocklists.
React Native
TypeScript
Claude API
Gemini API
OpenAI API
llama.cpp (on-device LLM)
Apple Foundation Models
SQLite
Zustand
Firebase
Disease Prediction Android Application
An Android app that predicts disease probability from user-entered symptoms. Built and trained a neural network in TensorFlow, improving prediction accuracy from 55% to 72% through iterative feature analysis, and optimized the preprocessing pipeline for a 30% efficiency gain.
Kotlin
TensorFlow
Neural Networks
Python
SQL

Experience

Software Developer

TechOwl InfoSec

March 2026 – Present
Remote
  • Develop and maintain cross-platform mobile applications using React Native, delivering consistent user experiences across Android and iOS.
  • Translate product requirements into functional mobile features following Agile development practices.
  • Integrate REST APIs and third-party services, connecting frontend interfaces with backend data sources.

Technical Support Specialist

Teleperformance

May 2024 – May 2026
Canada
  • Diagnosed and resolved complex software and application issues using structured troubleshooting methodologies, building systematic debugging skills applicable to software and AI application development.
  • Identified root causes of recurring technical issues and documented solutions, reinforcing analytical thinking directly relevant to debugging and improving software systems.
  • Mentored and onboarded new team members, translating complex technical processes into clear guidance.

Education

Post Graduate Diploma

Mobile Solutions Development

Conestoga College

GPA

3.46

Jan 2022 – Apr 2023
Waterloo, ON

Specialized in mobile application development with a focus on Android and iOS platforms. Covered advanced topics including UI/UX design, database management, mobile security, hybrid app frameworks, and full-stack integration through hands-on projects and a final capstone.

Bachelor Of Engineering

Computer Science Engineering

ITM Universe

Aug 2016 – Aug 2020
Vadodara, India

Comprehensive study of computer science fundamentals including algorithms, data structures, object-oriented programming, software engineering, networking, and databases—building a strong foundation for full-stack and system-level development.

Get In Touch

I'm always open to discussing new opportunities, interesting projects, or just having a chat about technology.

Contact Information

Email

bhoi.mihir7@gmail.com

Location

Gujarat, India

Availability

Open to full-time opportunities