Skip to content
User avatar
GitHub profile
LinkedIn profile

Manil Modi

GitHub profile
LinkedIn profile

@Manil09

Manil Modi

@Manil09

AI/ML engineer | scalable systems. Upcoming SWE Intern at Ignosis.

AI/ML engineer | scalable systems. Upcoming SWE Intern at Ignosis.

Skill iconPython
Machine Learning
Reinforcement Learning
Data Science
EDA

Nadiad, India

šŸš€ Manil Modi

AI/ML Engineer | Distributed Systems Builder | Agentic AI Developer

Upcoming Software Engineer Intern at Ignosis

I build production-grade AI systems that combine Machine Learning, Agentic AI, and scalable backend architectures.

My focus is not just training models but designing intelligent systems that operate reliably at scale.

šŸ§  Core Expertise

Machine Learning & AI

  • Deep Learning (CNN, RNN, LSTM)
  • Reinforcement Learning (PPO, RL agents)
  • Graph Neural Networks
  • NLP & LLM Systems
  • Time-Series Forecasting
  • Feature Engineering & Model Optimization

Agentic AI

  • Multi-Agent Systems
  • Agentic RAG
  • Tool-Using LLM Agents
  • Autonomous AI Pipelines
  • CrewAI Orchestration

Backend & System Design

  • Distributed Systems
  • Microservice Architecture
  • Event-Driven Systems
  • Transaction State Machines
  • Concurrency & Idempotency Design

MLOps & Infrastructure

  • Docker
  • AWS / GCP
  • Kafka Streaming
  • MLflow
  • Apache Airflow
  • Grafana & Prometheus Monitoring

āš” Key Projects

šŸ’° LedgerZero ā€” Intelligent Financial Infrastructure

Live: https://ledgerzero.xyz

Architected a financial transaction ecosystem with real-time fraud intelligence.

Highlights:

  • Transaction lifecycle state machine
  • Money laundering detection pipeline
  • Graph-based financial intelligence system

Architecture:

  • GNNs for suspicious transaction topology detection
  • RL agents for fraud pattern exploration
  • GraphRAG for investigative reasoning
  • Kafka streaming for transaction events

Tech Stack:

Spring Boot | FastAPI | Neo4j | PostgreSQL
Redis | Kafka | Docker | AWS
Graph Neural Networks | Reinforcement Learning

šŸ“ˆ Multi-Agent Stock Market Intelligence System

Built a multi-agent AI system that generates real-time financial reports using market data and technical indicators.

Features:

  • Live candlestick analysis
  • Financial ratio extraction
  • Market signal detection
  • AI-generated stock insights

ML Components:

  • XGBoost for OHLCV prediction
  • LSTM time-series forecasting
  • PPO reinforcement learning trading agent

Architecture:

  • CrewAI multi-agent orchestration
  • Real-time data ingestion
  • Agentic RAG for financial reasoning

Tech Stack:

CrewAI | FastAPI | MERN | WebSockets
XGBoost | LSTM | Plotly | Groq LLM

šŸ¤– Companion AI ā€” Intelligent Hiring System

Built an AI-powered recruitment pipeline that automates resume parsing, candidate scoring, and interview analysis.

Capabilities:

  • Resume entity extraction from PDFs
  • ATS-style candidate ranking
  • AI-powered interview evaluation

ML Features:

  • OCR-based skill extraction
  • Whisper speech-to-text
  • Video behavioral analysis with MediaPipe
  • Audio confidence analysis with Librosa

Tech Stack:

FastAPI | Dotnet Core
MediaPipe | OpenCV | Whisper
Transformers | Llama3

šŸ† Achievements

  • šŸ„‰ Top 3 ā€” DUHacks 5.0
  • šŸ„‡ Winner ā€” Holboxathon
  • AI/ML Lead ā€” Google Developer Groups DDU
  • Lead Organizer ā€” DUHacks 5.0

šŸ“« Connect With Me

GitHub: https://github.com/ManilModi
LinkedIn: https://www.linkedin.com/in/manil-modi-90b028278
Medium: https://medium.com/@msmodi1701

ā­ Always excited to collaborate on deep-tech AI systems and ambitious hackathon ideas