Gaurav
9,064 views
Gaurav Singh
location_onVellore, India

Gaurav Singh

ML Engineer & AI Researcher Amazon ML Summer School '26 Mentee Full-Stack Architect & Open Source Dev
codeGitHub workLinkedIn devicesGoogle Dev terminalLeetCode

About

school Academic Excellence

B.Tech Computer Science & Engineering at VIT Vellore (8.10 CGPA). Secured VITEEE Rank 19 (Top 0.1%) nationwide out of 2+ Lakh aspirants.

psychology Amazon ML Summer School

Selected among Top ~2% (~3,000 / 1.4 Lakh applicants). Completed 80+ hours of advanced modules mentored by Amazon Principal Scientists across DNNs, Transformers & Causal Inference.

  • • AI / ML & GenAI Architecture: Specialising in PyTorch, Graph Neural Networks (GNNs), LSTM Autoencoders, and production RAG / Agentic AI Workflows (LangChain, Hugging Face).
  • • High-Performance Systems & MLOps: Engineering resilient distributed systems using FastAPI, Redis, Docker, Kafka, and GitHub Actions CI/CD with sub-100ms real-time inference.
  • • Leadership & Competitive Track Record: Former Technical Head — GeoSpatial Club (mentoring 20+ researchers), 1st Place at NIT Durgapur Ideathon, Grand Finalist at Anveshana 3.0 & India Innovates, with 200+ DSA/SQL problems solved.

Certifications

verified
Google Associate Cloud Engineer (ACE)
Google Cloud · Certified
psychology
AWS Certified AI Practitioner
AIF-C01 · Amazon Web Services
smart_toy
IBM Agentic AI
IBM · Generative & Agent Systems
cloud
Oracle Cloud Infrastructure AI
2025 AI Foundations Associate · OCI

Featured Projects

Novellaio
Multi-Modal Generative AI Studio · Open-Source
MCP 2024-11-05 Ready
  • •Architected an interactive multi-modal AI storytelling platform powered by React 19, TypeScript 5.8, and Gemini 2.5, orchestrating dynamic CYOA branching trees, contextual visual art across 15 curated styles, and synchronized multi-lingual TTS narration.
  • •Engineered an integrated Model Context Protocol (MCP 2024-11-05) JSON-RPC 2.0 server exposing autonomous agent toolkits for Claude Desktop, Cursor IDE, and Gemini ADK.
  • •Implemented zero-token client-side publishing pipelines including EPUB 3.0 eBooks, worker-accelerated multi-theme PDF generation via OffscreenCanvas, and 9:16 / 16:9 social video reel exports with word-by-word karaoke subtitles.
React 19 TypeScript MCP Server Gemini 2.5 Web Audio API EPUB 3.0
Novellaio AI Story Studio
AEGIS.AI
Open-Source Project · Jul 2026 – Aug 2026
Sub-100ms Inference
  • •Architected an enterprise-grade behavioral anomaly detection engine utilizing PyTorch LSTM Autoencoders and Graph Neural Networks (GNN), achieving sub-100ms real-time inference latency (24.5ms P95).
  • •Engineered a multi-class threat classifier with an ensemble architecture and 5-Fold TimeSeries Split validation, securing 94.7% accuracy and a 91.8% PR-AUC on high-dimensional UEBA telemetry.
  • •Designed a resilient 4-tier fallback backend architecture for 99.9% availability, streaming real-time risk telemetry to a React SOC interface via FastAPI, Redis, and WebSockets.
PyTorch GNN LSTM Autoencoder FastAPI Redis WebSockets
AEGIS.AI Behavioral Threat Monitor
Bio-SentinelX
Capstone Research Project · Aug 2025 – Dec 2025
Winner at NIT Durgapur Hackathon
  • •Engineered production-grade Retrieval-Augmented Generation (RAG) and automated agent workflows using LangChain, ADK and Hugging Face maintaining 99.9% system uptime across 1,000+ daily vector queries.
  • •Architected end-to-end CI/CD pipelines using GitHub Actions to automate containerization and deployment, reducing manual release cycles by 60%.
  • •Developed high-throughput model serving pipelines using FastAPI and Docker, orchestrating real-time model inference under 150ms response time.
LangChain Hugging Face RAG FastAPI Docker GitHub Actions
Bio-SentinelX RAG Biomedical Engine
Arogya AI
Research Project · Jan 2025 – Apr 2025
399 ICD Conditions
  • •Designed a dual-engine architecture for multi-label disease prediction using Python, Scikit-learn, incorporating automated testing frameworks via PyTest to validate classifier robustness across 399 ICD-mapped conditions.
  • •Developed an 807-dimension TF-IDF NLP vectorization pipeline with high-precision cosine similarity matching, maintaining 98%+ prediction accuracy under peak query loads.
  • •Engineered secure Role-Based Access Control (RBAC) practitioner portals with integrated security checks, enforcing data privacy compliance for 500+ simulated clinical records.
Python Scikit-learn PyTest TF-IDF NLP Cosine Similarity RBAC
Arogya AI Clinical Diagnostic Pipeline
ShakeAlert
CDMM Research · May 2025 – Jul 2025
25% Accuracy Boost
  • •Built a full-stack earthquake early warning platform for India — low-latency seismic telemetry parsing with real-time Azure Maps coordinate dispatch and alerting.
  • •Integrated Firebase Firestore streaming pipelines with Pandas-based signal anomaly identification against live seismic sensor feeds from India's NDMA network.
  • •Implemented Python Seismology APIs to classify P-wave and S-wave arrivals, enabling sub-second alert dispatch to potentially affected regions.
Firebase Azure Maps Pandas Python
ShakeAlert Seismic Telemetry Platform

Experience

Mentee · Amazon ML Summer School
Jul 2026 — Aug 2026
Amazon Science · Remote
  • Selected into an elite group of ~3,000 mentees out of 1.4 Lakh national applicants (~2% acceptance rate) by clearing competitive multi-stage assessments in data structures, algorithms, and statistical aptitude.
  • Completed advanced training modules led by Amazon Principal Scientists across 80+ hours of curriculum across Deep Neural Networks, Sequence Modeling (LSTMs/Transformers), Probabilistic Graphical Models, and Causal Inference.
  • Validated machine learning competency via rigorous evaluations, mastering core optimization functions (Backpropagation, Gradient Descent), evaluation strategies (Bias-Variance tradeoff, PR-AUC), and production scaling constraints.
Research Intern
May 2025 — Jul 2025
Center Of Disaster Management & Mitigation (CDMM) · VIT Vellore
  • Architected a production-grade earthquake early-warning analytics platform, engineering an asynchronous data pipeline that eliminated manual geospatial assessment overhead.
  • Processed complex seismic signal datasets with automated preprocessing workflows, boosting trigger classification accuracy by 25% while mitigating false alarm rates.
  • Integrated 5+ disparate data ingestion feeds into a centralized dashboard architecture, refactoring backend services to seamlessly scale under shifting project requirements.
Technical Head
Mar 2025 — Mar 2026
VIT GeoSpatial Club · VIT Vellore
  • Directed technical operations for the "MapQuest" GIS treasure hunt during graVITas 2025, coordinating a core team to manage logistics for 100+ participants and a ₹9,000 prize pool.
  • Spearheaded collaborative workflows as Technical Head, overseeing the planning of 5+ technical workshops and mentoring 20+ junior members in advanced geospatial research and tool utilization.
  • Coordinated strategic event initiatives and cross-departmental logistics during the club's flagship activities, resulting in a 20% increase in overall project submission speed and engagement.

Education

B.Tech in Computer Science and Engineering
Vellore Institute of Technology (VIT) · Vellore, India
8.10 / 10 CGPA Class of 2027
VITEEE Rank 19 (Top 0.1% nationwide) Technical Head — GeoSpatial Club 200+ DSA & SQL Problems Solved
Key Coursework: Data Structures & Algorithms, Object-Oriented Programming, Operating Systems, Database Management Systems, Computer Networks, Machine Learning, Artificial Intelligence.
Deep learning framework — used in Bio-SentinelX ensemble & Arogya AI DenseNet training. Retrieval-Augmented Generation — powers Bio-SentinelX's dual-embedding medical search across 200M+ papers. LLM orchestration library — used for CoT reasoning chains and provider failover logic. Parameter-Efficient Fine-Tuning — enables fine-tuning large models with minimal compute. TensorFlow Lite — runs quantised Bio-SentinelX model directly inside the browser. Facebook AI Similarity Search — used for blazing-fast approximate nearest-neighbor retrieval.

Skills

psychology Machine Learning & Generative AI
PyTorch TensorFlow Scikit-learn LangChain Hugging Face RAG Agentic AI Deep Neural Networks Graph Neural Networks (GNN) LSTMs / Transformers Probabilistic Graphical Models Causal Inference SHAP / XAI TFLite
cloud_sync DevOps & MLOps
Docker Kubernetes GitHub Actions CI/CD Terraform Data Pipelines AWS (EC2, S3, Lambda) GCP (ACE)
code Languages & Frameworks
Python Java C/C++ SQL TypeScript FastAPI React Linux / Bash
database Databases & Streaming
Redis Kafka DynamoDB MySQL Supabase pgvector Firebase
groups Leadership & Collaboration
Technical Leadership Cross-Functional Collaboration Mentoring Analytical Problem Solving Strategic Event Coordination

Distinctions & Awards

Amazon ML Summer School 2026
Amazon Science · Selected in Top ~2% (~3,000 / 1.4 Lakh applicants) nationwide
Top ~2% · National CohortAug 2026
VITEEE Rank 19 (Top 0.1%)
Vellore Institute of Technology · National Engineering Entrance Examination
Rank 19VIT
IIC NIT Durgapur Ideathon 2026
Bio-SentinelX — Capstone RAG & Biomedical Platform
1st Place · 9.5/10May 2026
Anveshana 3.0 National Challenge
BMSIT&M, Bangalore · Top 50 nationwide
Grand FinalistMar 2026
India Innovates 2026
Municipal Corp of Delhi · Bharat Mandapam, New Delhi
Grand FinalistMar 2026
ET Gen AI Hackathon
Economic Times · Top Teams Nationwide
Phase 2 QualifierMar 2026

GitHub Activity

Contribution Calendar · github.com/gaur-avvv
View Profile →
GitHub contribution chart for gaur-avvv
GitHub streak stats
GitHub stats
Contact
Vellore Cable Station · Terminal 04
sensors
Despatch Nodes

Open to engineering roles · July 2026 · Vellore, India.