Recognition
Finalist — ANOKHA AI-Verse Hackathon
Amrita University
Pitched Agentic AI solutions for enterprise automation as a finalist in the ANOKHA AI-Verse Hackathon.
NIVEDH Manoj
AI / ML ENGINEERING STUDENT
Building explainable AI systems where thoughtful design meets applied intelligence.
Full-stack products across Graph RAG, medical imaging, low-resource NLP, voice analysis, and agentic workflows.
01 / About
About
I build applied AI systems that connect model reasoning with clear, usable interfaces.
My work moves across retrieval, medical imaging, language, speech, and agentic workflows, with an emphasis on making complex decisions inspectable.
- B.Tech CSE — AI & ML
- SRET / SRIHER
- Full-stack applied AI
- Graph RAG, vision & NLP
02 / Awards
Achievements
-
01 -
02 Leadership
Secretary — ML Club
–
Served as an MC for club events and helped host ideathons, AI quizzes, and debates.
-
03 Representation
Student Council
–
Facilitated student–faculty coordination across departments.
03 / Work
Selected Work
05 projects
Agx-RAG
An adaptive graph-augmented retrieval system for tracing evidence from indexed sources to an explainable answer.
Capability Runs a five-agent reasoning loop with contradiction-aware retrieval and local verification.
- TypeScript
- Graph RAG
- Agentic AI
RetinaScan AI
A multi-disease retinal screening workflow built around fundus imagery and interpretable model decisions.
Capability Explains retinal model predictions with Grad-CAM and SHAP visual attribution.
- Python
- EfficientNet-B2
- Grad-CAM
Approval Workflow System
An enterprise approval application that keeps creators, approvers, revisions, and budgets in one traceable path.
Capability Automates multi-stage, team-scoped approval routing with assisted validation.
- React
- Node.js
- OpenAI
Voice Analysis Assistant
An offline voice-therapy application that turns recorded speech into interpretable acoustic feedback.
Capability Extracts pitch, jitter, and shimmer, then adapts exercise difficulty with Q-learning.
- Flask
- Librosa
- Q-learning
Early Cancer Detection
A clinical risk-classification study using Wisconsin data for breast and lung cancer modelling.
Capability Represents probabilistic dependencies for interpretable risk prediction.
- Scikit-learn
- Bayesian network
- Clinical data
04 / Journey
Journey
Experience / Education
Trace 01
Experience
-
Agentic product development
AI/ML Intern
Developed AdSwift, a full-stack Agentic AI SaaS MVP that automates Google Ads creation using Next.js and Gemini.
Engineered a high-performance Python web-scraping pipeline using Crawl4AI and Playwright to extract structured business data in real time.
- Next.js
- Gemini
- Python
- Crawl4AI
- Playwright
- Agentic AI
-
Low-resource language research
AI/LLM Research Intern
Engineered a Transformer-based English–Tamil neural machine translation model from scratch with PyTorch for low-resource language translation.
Built a custom natural-language-processing pipeline for Named Entity Recognition and part-of-speech tagging without pretrained models.
- PyTorch
- Transformers
- English–Tamil NMT
- Named Entity Recognition
- POS Tagging
- Low-resource NLP
Trace 02
Education
-
Undergraduate
B.Tech CSE — AI & ML
Sri Ramachandra Faculty of Engineering and Technology (SRET)
Sri Ramachandra Institute of Higher Education and Research (SRIHER)
- Location
- Porur, Chennai, Tamil Nadu
- Period
- –
- CGPA
- 9.291 / 10
-
Class XII
Mahalashmi Vidhya Mandhir CBSE School
- Location
- Avadi, Tamil Nadu
- Board
- CBSE
-
Class X
Kulapati Munshi Bhavan's Vidya Mandir (KMBVM)
- Location
- Pottore, Thrissur, Kerala
05 / Stack
Stack
- languages
- Python, JavaScript, TypeScript, SQL, HTML/CSS
- frontend
- React, Next.js, Tailwind CSS, Vite, responsive UI systems
- backend
- Node.js, Express, FastAPI, Flask, REST APIs, JWT, bcrypt
- databases
- MongoDB, PostgreSQL, pgvector, vector search, knowledge graphs
- ai/ml
- PyTorch, TensorFlow, Scikit-learn, NLP, computer vision, RAG, MLOps
- tools and deployment
- GitHub Pages, Vite, REST API tooling, MLOps workflows
06 / Contact