Pune, India - AI Engineering

Aditya
Parbat

"Pre-final-year AI engineer building systems that coordinate, reason, and act autonomously - not just respond."

View Work Get Resume

Aditya Parbat / Profile Portrait

9.93GPA

#1RANK

2INTERNSHIPS

5+SHIPPED PROJECTS

98%BEST MODEL ACCURACY

Projects

LifeLink

4-agent autonomous blood donation coordination system

RESULT:< 2 MIN RESPONSE

PythonFastAPIGeminiCeleryRedisMongoDBADK

+4-agent system (AutoPulse, RapidAid, PathFinder, LinkBridge) for real-time inventory monitoring and emergency donor coordination
+Google ADK + Gemini AI powering LifeBot with A2A communication via MCP and full tool-trace visibility
+Async Celery + Redis orchestration across 8 blood groups, runs every 3 to 5 minutes

GitHub ↗

CogniFlow

Offline-first AI learning platform with local LLM inference

RESULT:100% OFFLINE - ZERO CLOUD

PythonTypeScriptOllamaRAGNext.jsSQLGemma3n

+Local LLM (Gemma3n via Ollama) for multilingual Q&A, auto-quizzes, and flashcards from uploaded PDFs
+React/Next.js dashboard with 7 learning games, Pomodoro timer, XP system
+On-device RAG pipeline for semantic indexing and accurate retrieval

GitHub ↗

AntiPiracy Engine

Real-time media piracy detection via computer vision

RESULT:98% DETECTION ACCURACY

PythonOpenCVMLDigital Watermarking

+CV feature extraction for real-time unauthorized media identification
+Digital watermarking for forensic attribution of pirated assets

GitHub ↗

Ctrl-A

Accessible welfare schemes platform for persons with disabilities

RESULT:PERSONALIZED SCHEME MATCHING

PythonFlaskFastAPIMongoDBChromaDBJavaScriptHTML/CSS

+Built a disability-first platform connecting PwD users to government welfare schemes through accessible search, voice navigation, and audio readout flows
+Implemented personalized scheme matching using disability type, state, age, and income filters with automatic eligibility-based ranking
+Designed a dual-backend architecture with Flask auth flows, FastAPI scheme APIs, MongoDB user data, and ChromaDB-powered semantic search

GitHub ↗

Experience

AI & CV INTERN

Cerebrospark Innovations Pvt Ltd

Jan 2026 - Present

"Built a ResNet50-based crop disease classifier achieving 93% recall and deployed a YOLOv11 surveillance model on 490+ annotated images with 5 object classes, optimized for edge inference. Also prototyped multi-agent RL simulations for swarm coordination."

AI INTERN

Goklyn Pvt Ltd

Nov 2025 - Feb 2026

"Engineered LLM pipelines via LangChain for production Q&A systems, architected a React + FastAPI + Docker containerized stack, and built a 4-agent validation pipeline processing 200+ records with parallel orchestration."