Deep Rathi
AI Engineer | Software & ML Developer
Building intelligent AI systems and scalable software.
About Me
Designing the future with AI and Code.
My Journey
I’m an AI Engineer at Key.ai with a passion for building production-grade AI systems that solve real-world problems. I specialize in generative AI, low-latency machine learning systems, and real-time optimization. With expertise in full-stack ML deployment and system design, I’ve built everything from autonomous driving models to high-frequency trading systems. My work focuses on bridging the gap between cutting-edge AI research and scalable production systems. Currently exploring the intersection of LLMs, real-time inference optimization, and infrastructure design.
📌 Key Highlights
Tech Stack
Tools and technologies I work with
Languages & Core
Machine Learning & AI
ML Frameworks & Libraries
System Design & Architecture
Web Development & DevOps
Specialized Domains
Experience
My professional journey
AI Engineer
Aug 2025 – Present- •Building and scaling AI-powered systems that drive the platform's core capabilities.
- •Developing intelligent features such as recommendation engines and other AI-driven solutions to enhance product experience.
- •Designing underlying infrastructure to ensure AI systems run reliably in production with focus on scalability, cost efficiency, and disaster recovery.
Machine Learning Intern
Feb 2025 – Jul 2025- •Built generative AI models for real-time inference, optimizing system pipelines for low-latency ML deployment.
- •Automated ML pipeline tasks like feature extraction, clustering, and hyperparameter tuning, improving production efficiency.
Data Scientist Intern
Aug 2023 – Oct 2023- •Developed interactive data dashboards (Tableau, Python, SQL) analyzing HR datasets for actionable business insights.
- •Collaborated with cross-functional teams to deliver data-driven solutions for operational improvements.
Research Intern
Jan 2022 – Oct 2022- •Contributed to LaneScan Net, a deep learning model for obstacle lane detection in autonomous driving.
- •Led data annotation, model training, and evaluation while collaborating with VIT and SUNY Binghamton researchers.
Featured Projects
A showcase of my recent work in AI, Machine Learning, and Software Development.
PrecisionEdge
AI-powered data analytics platform that automates 80% of data preprocessing using LLMs, reducing analytics time by 40%. Features intelligent data cleaning, automated feature engineering, and AI-driven insights generation. Streamlines the entire pipeline from raw data ingestion to actionable business intelligence. Published in AIP Scopus journal.
DeribitTradingSystem
Ultra-low-latency cryptocurrency trading system built in C++ for the Deribit exchange. Implements real-time order execution, market microstructure analysis, and advanced risk management. Achieves sub-millisecond latency for optimal trade execution. Features WebSocket-based market data handling and sophisticated order placement algorithms.
GROW (AI Learning Platform)
Intelligent personalized learning platform powered by GROQ LLM API. Generates custom study plans in under 5 seconds based on user goals and learning style. Includes progress tracking, adaptive difficulty adjustment, and exportable study materials. Leverages fast inference for real-time educational content generation.
Sad-Talker-Custom
Custom implementation of the SadTalker architecture for generating photorealistic talking head videos from audio input. Implements face synthesis, lip-sync generation, and head pose estimation. Combines diffusion models with expression transfer networks for high-quality video generation. Useful for video conferencing, content creation, and accessibility applications.
DSA Question Generator
Algorithmic system that generates unlimited data structures and algorithm problems with automatic validation. Features difficulty scaling, constraint generation, and solution verification. Includes comprehensive test case generation and performance analysis. Useful for interview preparation, competitive programming, and algorithm education.
SHL Recommender
Collaborative filtering-based recommendation engine with advanced feature engineering and model optimization. Implements matrix factorization, neural collaborative filtering, and hybrid recommendation approaches. Optimized for both accuracy and computational efficiency with production-ready data pipelines.
Achievements & Publications
Recognition and contributions to the field
AI-driven ad generation using Kolmogorov-Arnold networks
Published in AIP Scopus (Jan 7, 2026). Describes an AI-driven system for automated generation of personalized multimedia advertisements based on user interactions.
View Paper→PrecisionEdge: Cutting-edge data insights
Published in AIP Scopus (Jan 7, 2026). Open-source application designed to simplify and automate data analysis using advanced Large Language Models (LLMs).
View Paper→LaneScanNET: Deep-learning for autonomous driving
Published in Expert Systems with Applications (ScienceDirect). A deep-learning approach for simultaneous detection of obstacle-lane states for autonomous driving systems.
View Paper→Winner, IIM Ahmedabad AI Hackathon 2024
Secured 1st place for developing an innovative AI solution under time constraints.
Get In Touch
Have a project in mind or just want to say hi?