Computer Engineering graduate focused on building intelligent systems using Agentic AI, Generative AI, Large Language Models, Machine Learning, and Computer Vision.
I enjoy building AI systems end-to-end — from model experimentation and data processing to backend architecture, API integration, databases, and user-facing applications.
An AI-powered recruitment platform built around an adaptive, multi-panel agentic interviewer that conducts interactive voice interviews and evaluates candidates dynamically.
Momentum analyzes job requirements, company context, candidate information, and interview responses to determine:
- Interview panel composition and interviewer roles
- Evaluation criteria and required evidence
- Which interviewer should ask the next question
- Adaptive follow-up questions based on candidate responses
- Whether to continue, conclude, or initiate a system-design task
- Structured candidate assessment and evidence-based feedback
The platform combines unified agent reasoning, real-time voice interaction, adaptive interview stages, and system-design evaluation to deliver a more dynamic interview experience.
Tech: Python, FastAPI, Next.js, ElevenLabs, Groq, Gemini, Firebase, MongoDB
An AI-powered software engineering platform designed to investigate repository issues, identify root causes, implement code fixes, and validate changes through an agent-driven debugging workflow.
VectoraFix connects AI-assisted reasoning with repository-level engineering operations to help developers move from error diagnosis to tested code changes.
Key capabilities include:
- GitHub repository integration and authorized code access
- Repository inspection and issue investigation
- AI-assisted root cause analysis and debugging
- Automated code modification workflows
- Test-driven validation of proposed fixes
- MongoDB-backed application data management
- Backend APIs for coordinating repository repair operations
- A foundation for automating repetitive software maintenance tasks
Tech: Python, FastAPI, GitHub API, GitHub OAuth, MongoDB, AI Agents, LLMs, Next.js
An end-to-end AI pipeline for extracting and validating handwritten and printed information from Indian bank cheques.
The system combines specialized YOLOv8 detection models with OCR, confidence-based fallback recognition, handwritten text recognition, image preprocessing, and domain-specific validation.
Key components include:
- YOLOv8-based cheque field detection
- PaddleOCR primary recognition
- EasyOCR and TrOCR fallback recognition
- Transformer-based handwriting recognition
- OpenCV image preprocessing
- MICR validation
- Amount cross-verification
- Confidence-based OCR routing
- Streamlit-based processing interface
Tech: Python, YOLOv8, PyTorch, PaddleOCR, EasyOCR, TrOCR, OpenCV, Streamlit
An AI-driven decision-support system designed to improve energy supply chain resilience for import-dependent economies by combining geopolitical intelligence, supplier risk assessment, maritime route analysis, and LLM-based reasoning.
Key components include:
- Route Intelligence Engine
- Supplier Preference Engine
- Procurement Decision Engine
- Automated geopolitical news intelligence
- LLM-based structured insight extraction
- Dynamic supplier risk assessment
- Maritime chokepoint analysis
- Sanctions and diplomatic risk assessment
Tech: Python, FastAPI, Qwen2.5, Hugging Face Transformers, Pandas, Scikit-learn, NLP, LLMs
- Agentic AI
- Generative AI
- Large Language Models
- Multi-Agent Systems
- Natural Language Processing
- Computer Vision
- Voice AI
- Retrieval-Augmented Generation
- Vector Search
- Machine Learning
- Deep Learning
- AI Decision Systems
Languages & Core
Python, C++, JavaScript, TypeScript, SQL
AI / Machine Learning
PyTorch, TensorFlow, Scikit-learn, Hugging Face Transformers, Machine Learning, Deep Learning, NLP, Computer Vision, Object Detection, OCR, LLMs
Generative & Agentic AI
Agentic AI, Generative AI, LLM Applications, Multi-Agent Systems, RAG, Vector Search, AI Decision Systems
Backend & Databases
FastAPI, Flask, REST APIs, MongoDB, Firebase, Qdrant
Web Development
React.js, Next.js, Node.js, HTML5, CSS
Tools
OpenCV, Streamlit, Git
ChequeTrak: An Integrated Deep Learning Pipeline for Handwritten Text Extraction and Signature Region Extraction in Indian Bank Cheques
Published at the 2026 2nd International Conference on Computing, Communication and Green Engineering (CCGE).
- 2nd Prize — Vidyavardhini National Project Symposium (VNPS25)
- Semifinalist — HPCL PowerLab Innovation Challenge
- Top 34 teams out of 40,000+ participating teams
Software works registered with the Copyright Office, Government of India:
- Personalized Learning for Diverse Learners
- Personalized Learning App for Autistic Children
B.E. Computer Engineering
Vidyavardhini's College of Engineering & Technology
University of Mumbai
CGPA: 8.45 / 10
Email: kajaredhruv3@gmail.com