Applied AI Engineer · Bangalore, India

Generative AI & Agentic AI Engineer

Building reliable AI systems that retrieve, reason, act, and automate.

I work across prompt design, retrieval, tool calling, agent orchestration, evaluation, and deployment — with n8n handling the automation layer. Every agent I build runs against a clear success check, a cost budget, and a human-escalation path.

orchestration_pipeline
USER01CONTEXT02RETRIEVAL03REASONING04TOOLS05EVALUATION06ACTION07

About

Building AI Systems, Not Just AI Demos.

I'm an applied AI engineer focused on retrieval-augmented generation and autonomous agent workflows. My interest is in building AI applications where models have access to useful context, agents can use tools, and outputs are structured enough to act on.

I build evaluation and guardrails alongside features rather than after them. An AI workflow needs a measurable definition of success, defined boundaries, and a path for a human to step in when the system reaches the edge of what it should decide alone.

I'm comfortable owning the stack — prompt design and tool schemas, retrieval and vector storage, agent orchestration, and deployment — with n8n connecting those capabilities to real processes.

Useful contextTool accessStructured outputEvaluated systemsGuardrailsAutomatable workflowsHuman in the loop

Engineering Philosophy

How I Think About AI Systems

Five principles that shape how I design and ship AI systems.

01

Context First

Good AI applications depend on giving models the right context — retrieved, structured, and scoped to the task at hand.

Skills / Tech Stack

An Interactive Technology Ecosystem

AI engineering sits at the centre. Select a domain to see the technologies I work with there.

RAG

Retrieval-augmented generation over structured and unstructured sources.

  • RAG
  • Hybrid Retrieval
  • Context Engineering
  • pgvector
  • VectorDB
  • LangChain

Generative AI

Getting useful, structured output out of language models.

  • Context Engineering
  • RAG
  • Hybrid Retrieval
  • Structured Output
  • Fine-tuning
  • Evaluation
  • Guardrails

Agentic AI

Letting models plan, call tools, and coordinate.

  • Tool Calling
  • ReAct
  • Plan-Execute Loops
  • Multi-Agent Orchestration
  • n8n Workflow Orchestration

Frameworks

Libraries I reach for to compose AI systems.

  • LangChain
  • LangGraph
  • AutoGen
  • TensorFlow

Models & Serving

Models and runtimes I build against.

  • GPT-4o
  • Claude
  • Gemini
  • Llama 3
  • Mistral
  • Qwen
  • vLLM
  • Ollama

Data & Retrieval

Where context and memory actually live.

  • MySQL
  • Redis
  • PostgreSQL
  • VectorDB
  • pgvector

Programming

Languages I work in day to day.

  • Python
  • Core Java
  • JavaScript
  • SQL

Full Stack / Platform

Shipping the application around the model.

  • React.js
  • Node.js
  • FastAPI
  • Docker
  • Git
  • GitHub
  • Claude Code
  • Claude Cowork
  • Claude Design

Featured Projects

Systems, Pipelines, and Applications

Each project is presented as the pipeline it runs on. Open one for the problem, architecture, and engineering decisions.

01AI Building Design & Visualization Platform

PlotVision

Upload a plot image and get an automated read of its dimensions, orientation, and usable build area — then generate and refine building designs with natural-language prompts instead of CAD tooling.

ReactNode.jsPythonLLM APIs
flow
  • User Upload
  • Plot Image Analysis
  • Dimensions / Orientation
  • Buildable Area
  • Design Generation
  • Prompt-Based Editing
  • Version History
02Data Analytics

E-commerce Marketing & Sales Analysis

Cleaned and merged order, customer, and campaign data, then analysed sales across category, region, and channel to understand revenue contribution and repeat-purchase behaviour.

PythonPandasNumPyMatplotlibSeaborn
flow
  • Order / Customer / Campaign Data
  • Cleaning & Merging
  • Missing Values & Duplicates
  • Category / Region / Channel Analysis
  • Revenue & Repeat-Purchase
  • Visualization
03Machine Learning

ML Cryptocurrency Price Forecasting Platform

An ML platform that forecasts cryptocurrency prices with LSTM models trained on historical OHLCV data, served through a Django backend and a responsive frontend.

PythonTensorFlow/KerasLSTMDjangoREST APIsHTMLCSSJavaScript
flow
  • OHLCV Data
  • Data Cleaning
  • Normalization
  • Feature Engineering
  • LSTM
  • MSE / MAE
  • Forecast
  • Dashboard

Experience

Programs & Internships

  1. June 2026 – Present

    Scaler Academy

    AI/ML, Generative AI & Agentic AI Program

    Structured program covering AI/ML foundations, generative AI, and agentic AI systems.

  2. September 2025 – November 2025

    Infosys Springboard

    Artificial Intelligence Intern — Virtual Internship 6.0

    Virtual internship focused on artificial intelligence fundamentals and applied practice.

Engineering Experience

Owning the Stack Around the Model

Hands-on experience building the application layer for LLM systems — frontend, API, retrieval, tools, and evaluation.

application_architecture
  • Frontend
  • FastAPI
  • LLM
  • Retrieval
  • pgvector
  • Tools
  • Agent Orchestration
  • Evaluation

Frontend

  • React components for LLM applications
  • Chat interfaces
  • Streaming output
  • Prompt configuration panels

API & Orchestration

  • Python / FastAPI REST endpoints
  • Tool calling
  • Agent orchestration
  • Structured output parsing

Retrieval & Data

  • Retrieval pipelines
  • PostgreSQL and pgvector schemas
  • Document embeddings
  • Conversation history

Evaluation

  • Evaluation logs
  • Success checks
  • Cost budgets
  • Human-escalation paths

Represented as engineering capabilities, not employment positions.

Education

Academic Background

  1. 2022 – 2026

    B.Tech — Computer Science and Engineering

    Mohan Babu University

    CGPA 8.24 / 10
  2. 2020 – 2022

    Senior Secondary — Class XII

    Sree Vidyanikethan International School

    66%
  3. 2020

    SSC — Class X

    Narayana High School

    CGPA 9.0 / 10

Certifications

Job Simulations & Programs

AWS Solutions Architecture Job Simulation

Deloitte Data Analytics

Deloitte Cyber Security

Contact

Let's Build Something Intelligent.

I'm interested in building thoughtful AI systems, agentic workflows, and products where AI can create real utility.

Start a Conversation
Emailmaheshmaddina30@gmail.comLinkedInin/maheshmaddinaGitHubMaheshmaddina