Manish joshi

I'm a Software Engineer and 2025 Computer Science graduate focused on backend engineering, distributed systems, and AI-powered applications. I enjoy solving complex engineering problems, designing reliable systems, and taking products from initial ideas to working implementations.

I have 7 months of professional software development experience at Esferasoft Solutions, where I worked on production applications, integrated LLM-powered features into an AI fitness platform, and developed customer and driver applications with real-time tracking using Google Maps and WebSockets.

Alongside my professional experience, I've built several end-to-end engineering projects:

FlowForge – Distributed Workflow Execution Engine

Built a distributed workflow scheduling and execution system using Node.js, Express, Redis, BullMQ, MongoDB, and React. Implemented DAG processing, topological sorting, cycle detection, multi-worker execution, task leases, heartbeats, exponential backoff retries, and real-time monitoring. Added 100 automated Jest tests.

RepoGuard AI – Autonomous Code Analysis and Repair Platform

Developed an agentic code analysis system using Python, FastAPI, LLM tool calling, Python AST, and sandboxed PyTest execution. Built an eight-stage pipeline for identifying code issues, validating findings, and generating GitHub pull requests. Implemented multi-provider LLM fallback and 94 automated tests.

SpeakFlow – AI English Coaching Application

Built and published an AI-powered English-speaking application on Google Play using Flutter, LiveKit WebRTC, Gemini, and Supabase. Implemented real-time AI voice conversations, pronunciation evaluation, user progress tracking, and a structured 30-day learning curriculum.

My primary technical interests include backend architecture, distributed systems, API design, asynchronous processing, AI agents, RAG systems, LLM applications, and machine learning.

I'm particularly interested in Software Engineer, Backend Engineer, Full Stack Engineer, and AI Engineer opportunities where I can solve challenging problems, take ownership of features, and build reliable products.