Atharva Kanawade
From multi-agent AI systems and production backends to mobile apps and racing drones — I engineer solutions across the full stack. Currently building AI-driven production systems at Applied AI Consulting, and always shipping something new.
About Me
Background
Electronics & Telecommunication Engineering student with professional experience in AI-driven production systems. Currently completing a 6-month industry internship at Applied AI Consulting, building multi-agent AI workflows and production backends while maintaining a 9.02 CGPA.
What I Build
I work across the entire stack — from agentic AI systems and LLM integrations to mobile apps (React Native), full-stack backends (FastAPI, Node.js), and even racing drones. Every project ships production-ready, not as a side experiment.
Engineering Philosophy
Security thinking is baked into everything I build. Like a certain genius billionaire — I design, I build, I test, I iterate. From the first line of code to deployment, every system is engineered to be production-grade.
Professional Experience
Software Engineer – Trainee
Applied AI Consulting (AAIC)
Worked on AI-driven production systems and backend workflows as part of a 6-month industry internship.
Key Contributions
- ▹Contributed to multi-agent AI system workflows and orchestration
- ▹Improved test coverage to ~70% across multiple modules
- ▹Validated SQL to PostgreSQL migration including tables, indexes, constraints, and procedures
- ▹Investigated live production issues using AWS CloudWatch and RDS logs
- ▹Worked with LangGraph, WebSockets, FastAPI, and AI models such as OpenAI Whisper and Gemini 2.5
- ▹Contributed to React Native application development
- ▹Gained hands-on exposure to webhook systems, queue workers, and log-based debugging
Focus: Agentic AI systems, backend engineering, cloud debugging, and system reliability.
🍳 Cooking Application (During Internship)
LLM and database heavy cooking application developed alongside the internship.
- ▹Full-stack mobile app with LLM-powered recipe intelligence
- ▹Firebase for authentication and App Check
- ▹Global state management via Redux Toolkit
- ▹Hosted on AWS, developed for both iOS and Android
R&D Intern
Encrypta Inc.
Key Contributions
- ▹Engineered secure desktop authenticator using Python, Node.js, and Electron.js
- ▹Implemented HTTPS and WebSocket encryption achieving zero data leaks in testing
- ▹Deployed multi-factor authentication reducing attack vectors by 70%
- ▹Applied secure coding practices throughout the development lifecycle
AI-ML Head – UAV & Drone Projects
Team Vajra, MMCOE
Key Contributions
- ▹Managed end-to-end UAV development using Pixhawk, APM, and SpeedyBee flight controllers
- ▹Supervised flight tuning, testing, and integration of embedded and automation systems
- ▹Developed a compact 3-inch FPV racing drone optimized for confined-space flying
- ▹Designed and tuned a high-performance 5-inch FPV racing drone for competitive racing
- ▹Led AI/Autonomous navigation research for UAV applications
Featured Projects
From mobile apps and AI systems to drones and security platforms — here's what I've built.
Cooking Application
FeaturedLLM-Powered Recipe Intelligence Platform (Built during AAIC Internship)
Full-stack mobile cooking application leveraging LLM intelligence for recipe discovery, built as a side project during my internship at AAIC.
- ▹LLM-powered recipe search and recommendation engine
- ▹Firebase Auth + App Check for secure authentication
- ▹Redux Toolkit for global state management
- ▹Hosted on AWS, developed for both iOS and Android
Mobile app for precursor chemical supply chain management with ML anomaly detection, IoT sensor integration, and blockchain traceability.
- ▹End-to-end supply chain tracking mobile application
- ▹LSTM and Isolation Forest ML models for anomaly detection
- ▹IoT sensors connected via ThingSpeak for real-time monitoring
- ▹Blockchain scripts in Python for traceability
Ad-free Indian music streaming app for Android. Streams audio from YouTube via yt-dlp with zero cloud storage and zero infra cost.
- ▹Streams audio from YouTube via yt-dlp — no cloud storage needed
- ▹Supabase Auth (JWT) and PostgreSQL database
- ▹Zustand + React Query for state management
- ▹react-native-track-player v4 for audio playback
Mobile task management app where users create priority-based tasks, pin them to the home screen, and track efficiency via a dashboard.
- ▹Priority-based task creation and management
- ▹Pin tasks directly to device home screen
- ▹Efficiency tracking dashboard
- ▹Supabase for auth and database backend
Lightweight computer vision project for classifying Capacitors and Inductors using MobileNetV2 transfer learning, designed for future edge deployment.
- ▹MobileNetV2 transfer learning — 99-100% validation accuracy
- ▹Binary classification: Capacitor vs Inductor
- ▹Designed to be lightweight for future ESP32-CAM deployment
- ▹TFLite conversion planned for edge inference
Data-driven study of IPL match predictability using only pre-match information. Features a confidence-aware CatBoost model that knows when to abstain.
- ▹Confidence-aware model that abstains on low-signal matches
- ▹66.7% accuracy on high-confidence predictions (IPL 2024)
- ▹Symmetric team labelling to remove positional bias
- ▹Final deployment model trained on 2019-2025 data for IPL 2026
Full-spectrum security platform combining ML-powered phishing detection with hardware-based wireless defense and cloud honeypot infrastructure.
- ▹Naive Bayes classifier achieving 90%+ phishing detection accuracy
- ▹ESP8266/ESP32 system blocking 85%+ deauth attacks
- ▹AWS EC2 honeypots for threat intelligence gathering
- ▹Lightweight WAF with SQLi/XSS protection and rate limiting
IoT-enabled agricultural rover using CNN-based plant disease detection with cloud updates and automated farmer alerts.
- ▹CNN model with 90%+ accuracy for potato leaf disease detection
- ▹Custom rover with sensors for autonomous field monitoring
- ▹Real-time cloud data transmission via ThingSpeak
- ▹SMS alert system using Twilio API integration
Enterprise-grade React application that automates NIST SP 800-61 compliant incident response plans.
- ▹NIST SP 800-61 compliant plan generation
- ▹90% reduction in manual IR planning time
- ▹PDF export in under 2 minutes
- ▹Validated by 92% of enterprise test users
Automotive Part QC Classifier
Industrial AI Quality Control System
Production-ready computer vision system for automated part inspection using transfer learning.
- ▹Transfer learning with MobileNetV2 for part classification
- ▹Reliable detection across multiple part variants
- ▹Packaged as .exe for factory floor deployment
- ▹Real-time defect detection pipeline
Autonomous quadcopter for short-range payload delivery using CrossFlight FC and custom low-cost telemetry module.
- ▹Fully autonomous waypoint-driven flight system
- ▹Custom low-cost telemetry module with stable real-time link
- ▹CrossFlight FC integration for precision navigation
- ▹Designed for short-range delivery missions
Lightweight 3-inch FPV drone built using SpeedyBee Mini FC, optimized for confined-space flying with refined PID loops.
- ▹Compact 3-inch FPV platform for tight-space flying
- ▹Custom PID tuning for responsiveness and stability
- ▹Optimized for power efficiency
- ▹Lightweight frame for agile and precise flight
High-speed 5-inch FPV racing drone developed using SpeedyBee F405 V3 FC, optimized for competitive racing.
- ▹High-performance racing drone for competitive events
- ▹Precision-tuned for speed and sharp cornering
- ▹SpeedyBee F405 V3 for reliable control
- ▹Optimized flight stack for aggressive maneuvers
Technical Arsenal
Tools and technologies I use to build production systems across the full stack.
Agentic AI & LLMs
Multi-Agent Systems
ADVANCEDLLM Integration
ADVANCEDAI Backend Systems
STRONGMobile Development
React Native
ADVANCEDState Management
STRONGMobile Backend
STRONGMachine Learning
Deep Learning
ADVANCEDClassical ML
ADVANCEDComputer Vision
STRONGModel Deployment
STRONGSecurity
Penetration Testing
ADVANCEDNetwork Security
ADVANCEDThreat Detection
STRONGSecure Development
STRONGEngineering
Backend Development
ADVANCEDFrontend Development
STRONGCloud & DevOps
STRONGDatabase Systems
STRONGRobotics
Flight Controllers
ADVANCEDAutonomous Navigation
STRONGFPV Tuning
ADVANCEDDrone Systems
EXPERTAdditional Technologies
Achievements & Certifications
Achievements
ARTEMIS Published in IEEE Xplore
Presented at 3rd IEEE International Conference on Emerging Trends and Innovation in ICT (ICEI 2026), PICT Pune
View on IEEE XploreWorld Rank 3 & AIR 2 – Technoxian Drone Racing 2025
Ranked among 32 international teams in global FPV drone racing competition
CTF Winner - NigVanta'25
Capture The Flag Competition Champion
Innovators Challenge 2k24
24-Hour Hackathon Winners
Certifications
Cisco Certified Network Associate (CCNA)
Cisco • May 2025
Google Cybersecurity Professional Certificate
Google & Coursera • Jul 2024
Network Essentials
Cisco • Jun 2024
From the Blog
Let's Connect
Interested in collaborating on AI systems, full-stack projects, mobile apps, or anything engineering? I'm always open to new opportunities.