Hello Everyone
I'm Rohit Mishra
01 THE STORY SO FAR
Hello! I’m Rohit Mishra, a Software & Applied AI Engineer based in Indore, India, with 2+ years of experience architecting intelligent backend systems and autonomous workflows.
Holding a degree in Artificial Intelligence & Data Science, I’ve translated deep technical foundations into high-impact production applications—designing multi-agent architectures, event-driven pipelines, and robust web platforms that solve complex operational challenges.
I have always been fascinated by how technology can be leveraged to solve real-world problems. When I first saw how modern AI could move beyond static data processing to actively converse, reason, and orchestrate decisions, I was completely hooked. For me, it isn't just about writing code—it's about giving software a brain.
Over the past couple of years, I’ve worked closely with fast-moving teams and startups, bridging the gap between cutting-edge LLMs and scalable backend infrastructure to build digital experiences that feel intelligent, resilient, and genuinely helpful.
When I’m not orchestrating agentic workflows or designing distributed systems, you can usually find me diving into the latest open-source AI tech, reading sci-fi, or playing chess.
02 WHAT I BUILD
In my day-to-day work, my main focus is bridging the gap between raw AI capabilities and practical human needs.
I spend a lot of my time automating the mundane. I build systems that take over repetitive tasks—like creating a voice-to-voice AI agent that can conduct initial job interviews over Google Meet, or designing automated pipelines that instantly generate dynamic price quotes from a simple WhatsApp text. My goal is always to make software do the heavy lifting so people can focus on the creative side of their work.
Beyond automation, I also love building from the ground up. Having worked as a founding developer for startups, I enjoy crafting full-stack platforms, from designing the initial user interface to setting up real-time chat servers and databases. Whether I am engineering an 'AI Brain' for a personalized learning platform or putting together a sleek e-commerce site, I handle the architecture, the logic, and the final polish.
03 HOW I COLLABORATE
I’d much rather be a true partner in a project than just an executor who blindly writes code from a to-do list.
I thrive in startup-like environments where communication is open, casual, and highly collaborative. I love being involved from the very beginning—understanding the core of the problem, brainstorming the user experience, and figuring out the best way to bring an idea to life.
Instead of rigid, formal processes, I prefer jumping into quick conversations, sketching out ideas, and building rapid prototypes to test if a concept works before we commit to the heavy engineering. I believe that the best products are built when everyone on the team is aligned, comfortable, and working together seamlessly.
Current Playground
I'm architecting context-aware 'AI Brains' powered by distributed memory systems or engineering multi-agent networks that autonomously reason through complex research, I am constantly pushing the boundaries of intelligent software. I treat my personal projects like production-grade applications—focusing on modular architecture, dynamic prompt engineering, and real-world utility rather than simple scripts. You can explore my latest AI workflows, complex automation engines, and full-stack tools over on my GitHub.
Work
A selection of my experiences where I've applied my skills in AI and engineering.
Applied AI & Backend Engineer
Oct 2025 - Present- Architected a production AI automation platform that transformed unstructured travel operations into optimized fleet schedules using FastAPI, React, Redis, and LLM-driven document understanding.
- Built an optimization engine for vehicle scheduling under capacity, duty-hour, and operational constraints, minimizing fleet usage while generating fully auditable planning decisions.
- Developed secure payment middleware integrating Indecab and Cashfree with real-time verification, OTP authorization, webhook processing, idempotency, and financial audit trails.
- Engineered an AI-powered quotation engine combining LLM reasoning with deterministic pricing algorithms, Google Maps routing, and fuzzy entity matching for automated travel quotations.
- Designed event-driven AI automation workflows using n8n, OpenAI, Recall.ai, Bolna.ai, Unipile, and Google Workspace to automate recruitment, sales outreach, logistics, and voice operations.
FastAPIReactRedisn8nOpenAIRecall.aiBolna.aiUnipileAI Backend Engineer
Apr 2025 - Sep 2025- Engineered the backend gateway for a gamified AI learning platform using Node.js, Express, Redis, Firestore, and Google Gemini, enabling reliable lesson delivery and real-time student progress tracking.
- Built the Spacey Nova mentor architecture with intent classification, task routing, and specialized handlers for doubt resolution, guidance, encouragement, and general conversation using LangChain and LangGraph.
- Developed the Socratic lesson evaluation engine using structured prompts, low-temperature Gemini inference, and an auto-repair JSON validation loop, reducing misclassification of student responses.
- Implemented a Redis-backed conversation memory system with an in-memory fallback, ensuring persistent, thread-isolated tutoring sessions and lower response latency.
- Coordinated the real-time voice interaction pipeline — session lifecycle, speech synthesis, and turn-based conversation — using Supabase for session state and WebSocket streaming.
Node.jsExpressRedisFirestoreGeminiLangChainLangGraphSupabaseFounding Software Engineer
Oct 2024 - Mar 2025- Built and deployed multiple production-grade web applications — corporate websites, e-commerce platforms, real-time communication systems, and portfolio sites — using React, Node.js, Flask, MongoDB, and Railway.
- Architected full-stack application backends with authentication, REST APIs, database design, and deployment pipelines, enabling scalable client-facing products across multiple business domains.
- Engineered a real-time messaging platform using Flask, MongoDB, and WebSockets, delivering low-latency bidirectional communication and persistent chat functionality.
- Led frontend development using React and Tailwind CSS, translating product requirements into responsive, high-performance interfaces with a reusable component architecture.
- Established collaborative engineering workflows using GitHub, Postman, and code reviews to improve API quality and accelerate feature delivery.
ReactNode.jsFlaskMongoDBRailwayWebSocketsPostmanData Analyst Associate
Jul 2024 - Sep 2024- Designed end-to-end data analytics workflows using Python, SQL, PostgreSQL, and Looker Studio, transforming raw datasets into actionable business intelligence.
- Built scalable data preprocessing pipelines to clean, standardize, and validate large datasets, improving downstream analytical accuracy and reporting reliability.
- Engineered normalized PostgreSQL database schemas and optimized query performance to support analytical dashboards and recurring reporting workloads.
- Developed interactive dashboards and visual analytics that enabled stakeholders to monitor key business metrics and make data-driven decisions.
- Performed exploratory data analysis and statistical investigation to identify trends, anomalies, and data-quality issues that informed business recommendations.
PythonSQLPostgreSQLLooker Studio
My Projects
Check out some of my recent work below.

Vantage
- Engineered a multi-agentic stock analysis pipeline using LangGraph and FastAPI to automate equity research, sentiment analysis, and quantitative risk assessment for real-time ticker data.
- Implemented a stateful graph architecture with specialized nodes for news scraping (yfinance), sentiment scoring, and volatility calculations, using AgentState as a single source of truth.
- Integrated a hybrid LLM strategy supporting local inference via Ollama (fine-tuned GGUF models) and cloud-based providers (Gemini, Groq) with JSON-mode enforcement for consistent state merging.
- Developed a high-performance React Terminal UI with Framer Motion micro-animations, Vite builds, and a responsive design for streaming agent updates.
- Deployed a full-stack infrastructure on Railway (backend), Vercel (frontend), and Supabase (JWT-based auth) for secure, scalable access to the research pipeline.
- Architected a quantitative risk engine calculating annualized volatility and daily log returns, automating Buy/Hold/Sell decisions from configurable sentiment and risk thresholds.

Garment
- Developed a high-performance e-commerce storefront using React 18 and Vite, featuring complex state management for cart and checkout.
- Optimized SPA routing and load times via React Router v6, lazy loading, and intelligent data caching strategies.
- Integrated Stripe for secure payments and implemented robust schema validation using Formik and Yup.
- Leveraged Framer Motion for premium animations and React Window virtualization for efficient product rendering.

Deep Research Agent
- Built an autonomous AI research assistant leveraging LLM reasoning, RAG, and multi-agent orchestration for complex query resolution.
- Developed a dual-agent LangChain system to ingest and analyze contextual data from web and document sources via Tavily and Cohere APIs.
- Architected a Streamlit interface for real-time interaction, integrated with DeepEval for rigorous output quality benchmarking.
- Engineered a modular, scalable framework designed for extensibility and advanced multi-agent support.

Email Spam Classifier
- Built a machine learning model using NLP techniques to classify emails as spam or not spam.
- Applied text preprocessing steps including tokenization, stopword removal, and TF-IDF vectorization using NLTK.
- Trained and evaluated multiple Naive Bayes classifiers; achieved 97.85% accuracy using Bernoulli NB with hyperparameter tuning.
- Deployed the model with a user-friendly interface using Streamlit.
Contact Me
Use the form or links below.