AI Engineer

Leeon John

Building intelligent systems
that actually work.

AI Engineer focused on building production-grade LLM, agentic AI, and ML systems. Experienced in RAG, semantic retrieval, LLM evaluation, observability, AI APIs, and scalable Python backends.

Enter the system ↓GitHub ↗LinkedIn ↗

How the work flows

INPUTSEE: WORKRETRIEVALSEE: CAPABILITIESREASONINGSEE: WORKEXECUTIONSEE: WORKOBSERVABILITYSEE: WORK
  1. 01Input
  2. 02Retrieval
  3. 03Reasoning
  4. 04Execution
  5. 05Observability

Index

One environment, five stations.

The pipeline above resolves into this map. Every station is wired to the same core — choose one to travel to that part of the site.

Work

Systems, not screenshots.

Four production-oriented builds from the resume. Each one runs — send a trace, run a benchmark, analyze an issue.

Project 01 — API → auth → service → repository → isolated store

Omnira

LLM Observability Platform

  • Built an LLM observability platform using Python, FastAPI, SQLAlchemy Async, Pydantic, and Hexagonal Architecture.
  • Implemented authenticated trace ingestion and querying with pagination, advanced filtering, search, repository abstractions, and project-level isolation.

Stack · Python · FastAPI · SQLAlchemy Async · Pydantic · Hexagonal Architecture

Live system

Fig. 01

Loading interactive demonstration…

Project 02 — planner → discovery → analysis → retrieval → cited report

ResearchPilot-MCP

Multi-Agent Research Assistant

  • Built a multi-agent research system using MCP, LangGraph, FastAPI, and ChromaDB for paper discovery, PDF analysis, semantic retrieval, and citation-aware reports.
  • Designed agent workflows for research planning, document analysis, retrieval, and automated report generation.

Stack · MCP · LangGraph · FastAPI · ChromaDB

Live system

Fig. 02

Loading interactive demonstration…

Project 03 — run → measure → optimize → fall back

BenchLytics

LLM Evaluation & Inference Optimization

  • Built an LLM evaluation system for measuring model quality, latency, cost, failures, and inference performance.
  • Implemented asynchronous execution, dynamic batching, multi-tier caching, evaluation pipelines, and fallback strategies.

Stack · Async execution · Dynamic batching · Multi-tier caching · Evaluation pipelines

Live system

Fig. 03

Loading interactive demonstration…

Project 04 — repo → semantic index → plan → tests → PR draft

OpenSourcePilot

AI-Powered Open Source Assistant

  • Built an AI system that analyzes GitHub repositories and generates issue-specific contribution plans using LLMs and semantic code search.
  • Implemented ChromaDB retrieval, transformer embeddings, automated test generation, and pull-request drafting through GitHub APIs.

Stack · LLMs · ChromaDB · Transformer embeddings · GitHub APIs

Live system

Fig. 04

Loading interactive demonstration…

Research

Accepted, not aspirational.

Three peer-accepted papers. Only what the record states — expand a row for the citation.

Citation

Paper accepted at IEEE ICCCNT 2025 (IIT Indore). Leeon John.

Experience

Two roles, both hands-on.

Internships where pipelines and models shipped — stated exactly as they happened.

  1. AI Development Intern — Schbang Solutions

    • Built scalable ML pipelines for automation and customer insight extraction.
    • Developed GenAI-driven content solutions using transformer-based models.

    Jun 2025 – Sep 2025

  2. AI/ML Intern — Bits Infotech

    • Developed recommendation systems and predictive ML pipelines for e-commerce data.
    • Built supervised learning workflows using Python and scikit-learn.

    Jan 2024 – May 2024

Capabilities

A capability map, not a wall.

Five categories from the resume. Select one to read its register — then select a technology to trace its connections.

AI & LLM

Where the actual product work happens.

Select a technology to see its connections.

About

Concise, like a good log.

Who this is, where the degrees are from, and how to reach him.

About

AI Engineer building production-grade LLM and agentic systems — retrieval, evaluation, and observability included, not bolted on.

leeonjohn.work@gmail.com

+91-9499789565

github.com/coderleeon ↗

linkedin.com/in/leeon-john ↗

Education

  1. Pandit Deendayal Energy University (PDEU)

    M.Tech in Artificial Intelligence

    Aug 2024 – May 2026

  2. Parul University

    B.Tech in Computer Science and Engineering

    Jul 2020 – May 2024

Open to AI engineering roles where evaluation, retrieval quality, and observability are first-class requirements.

© 2026 Leeon John

Set in Newsreader & Inter