sairaj@kernel-workstation:~$ neofetch
-----------------------------------
[OS]         : macOS / Arch Linux (POSIX Compliant)
[Identity]   : SAIRAJ MOULASAB NADAF
[Role]       : Autonomous AI Systems Architect | Full-Stack & Systems Engineer
[Education]  : Master of Computer Applications (MCA) @ SMVIT College, Bangalore
               Bachelor of Computer Applications (BCA) @ SME's Abdul Kalam College
[Core Tech]  : Multi-Agent Runtimes, Spring Boot 3.3, FastAPI, Next.js 14, PostGIS, Solidity
[Leadership] : Founder & Tech Lead @ Mindrift (Sir MVIT AI Technology Club)
[Credentials]: ISC2 Certified in Cybersecurity (CC) | NVIDIA Dev Member | AMD Dev | Cyfrin | n8n
[Telemetry]  : 100+ Day GitHub Commit Streak 🔥 | 20+ Merged PRs 🔀 | 6+ Hackathon Ships 🏆
[Status]     : Seeking Software Engineering / AI Systems Internships

Computer Applications graduate student (MCA @ SMVIT Bangalore) and Systems/AI Engineer with 3+ years of experience architecting production-grade software across Agentic AI, cloud infrastructure, data engineering, application security, and Web3 protocols.

• Autonomous Agentic Runtimes: Designed bounded-autonomy multi-agent DAG & event-driven runtimes utilizing Google Vertex AI ADK, Gemini 2.0/Claude APIs, AsyncIO, and Pydantic structured schemas with SHA-256 state machines and refusal-recording audit trails.
• High-Performance Distributed Backends: Built resilient backend microservices with Java 21, Spring Boot 3.3, FastAPI, Node.js, PostgreSQL/PostGIS, Redis, and row-level concurrency controls.
• Geospatial Data Engineering & ML: Developed real-time geospatial risk prediction platforms using Smile Gradient Boosting, 5+ API ingestion streams on 15-min cycles, and H3 hexagonal spatial indexing.
• Security-First Architecture: Certified by ISC2 in Cybersecurity (CC). Implemented Row-Level Security (RLS), OAuth 2.0, JWT authentication, EIP-712 gasless transactions, and capability-scoped execution envelopes.
• Community & Hackathons: Founder and Lead of Mindrift AI Club at Sir MVIT. Shipped 6+ hackathon projects including ETHGlobal and Hack Odisha.

 ┌────────────────────────────────────────────────────────────────────────────────────────┐
 │                                 ENGINEERING DOMAINS                                   │
 ├───────────────────┬───────────────────┬───────────────────┬────────────────────────────┤
 │  AI & AGENTIC     │  BACKEND & CLOUD  │ DATA & GEOSPATIAL │   SECURITY & BLOCKCHAIN    │
 ├───────────────────┼───────────────────┼───────────────────┼────────────────────────────┤
 │ • Vertex AI ADK   │ • Java 21 / Spring│ • PostGIS / Spatial│ • ISC2 CC Security         │
 │ • Gemini & Claude │ • Python AsyncIO  │ • pgvector (RAG)  │ • OAuth 2.0 / JWT          │
 │ • Multi-Agent DAGs│ • FastAPI / Node  │ • H3 Hex Grids    │ • Solidity / Foundry       │
 │ • Event-Sourcing  │ • Next.js 14 App  │ • Redis / PubSub  │ • EIP-712 Gasless          │
 │ • Pydantic Output │ • Docker / GCP    │ • Smile ML (GBDT) │ • Audit Log Refusals       │
 └───────────────────┴───────────────────┴───────────────────┴────────────────────────────┘

Tech Stack: Python • Google Vertex AI Agents (ADK) • Gemini API • Cloud Run • Firestore • Gmail API • OAuth 2.0 • Pydantic

flowchart TD
    A[Gmail Pub/Sub Triggers] --> B[FastAPI Backend / Cloud Run]
    B --> C[Vertex AI Agent Pipeline - ADK]
    
    subgraph ADK ["Google Vertex AI Agent Pipeline (ADK)"]
        D["1. Extraction Agent"] --> E["2. Anomaly Evaluation"]
        E --> F["3. Dispute Drafting"]
        F --> G["4. Human Approval Gate"]
        G --> H["5. Follow-up Agent"]
    end
    
    C --> I[("Firestore State Machine (SHA-256 Case Keys & Token Accounting)")]

• Architected a 5-agent LLM pipeline (Extraction → Anomaly Analysis → Dispute Drafting → Human-Approval Gate → Follow-up) deployed on Google Vertex AI Agent Development Kit.
• Enforced deterministic structured JSON output using strict Pydantic validation schemas.
• Designed an idempotent Firestore state machine utilizing SHA-256 case keys, state transition logs, and per-call token accounting.
• Deployed as a serverless FastAPI microservice on Google Cloud Run integrated with Gmail Pub/Sub triggers and OAuth 2.0 authentication.

Tech Stack: Python AsyncIO • Gemini API • Event-Sourcing • OpenTelemetry • Capability Security

flowchart TD
    A["Event Stream (Cron / Webhook / PubSub)"] --> B["AsyncIO Runtime Loop"]
    B --> C{"Capability Check\n(Owner / Untrusted)"}
    C -->|Approved| D["Dynamic Task Frontier Proposal"]
    D --> E["Windowed Token Budget & Concurrency Lease Lock"]
    E --> F["OpenTelemetry Refusal Audit Trail"]

• Built a general-purpose agentic runtime that plans, executes, and critiques evidence in a continuous loop, dynamically proposing runnable task frontiers rather than relying on a static DAG.
• Autonomy Governance Engine: Implemented windowed token budgets, lease-based concurrency locks, and refusal-recording audit trails.
• Verified execution safety via a reproducible proof suite covering seven bounded-autonomy properties.
• Normalized cron, webhook, Pub/Sub, and multi-channel (Discord/Gmail/Slack/Telegram) events into a canonical envelope with an owner/user/untrusted capability security model.

Tech Stack: Java 21 • Spring Boot 3.3 • PostgreSQL / PostGIS • Redis • Smile ML (Gradient Boosting) • Leaflet.js

flowchart LR
    A["5x Weather & Climate APIs"] -- "15-min cycle" --> B["PostGIS Ingestion Engine"]
    B --> C["Smile Gradient Boosting ML Model"]
    C --> D["H3 Hexagonal Grid (0-100 Risk)"]
    D --> E["Spring Boot 3.3 REST API & Leaflet.js"]

• Engineered a full-stack geospatial machine learning platform predicting street-level flood and heat risk (0–100 risk index) across H3 hexagonal city grid cells.
• Integrated real-time data ingestion pipelines querying 5 weather & climate APIs on a 15-minute cron cycle.
• Trained gradient boosting models (using Java's Smile ML library) on 5+ years of historical climate data with an explainability layer highlighting top risk vectors.
• Secured backend endpoints via JWT authentication, RBAC, and spatial queries optimized with PostGIS GIST indexes.

Tech Stack: Next.js 14 (App Router) • TypeScript • Supabase (PostgreSQL) • Zustand • React Hook Form • Zod

• Built a real-time classroom and campus resource scheduling engine with atomic, race-condition-free conflict detection implemented via PostgreSQL RPC transactions.
• Enforced fine-grained Row-Level Security (RLS) policies for student, faculty, and administrator data access.
• Integrated Supabase Realtime for instant slot updates, automated approval workflows, and an administrative utilization dashboard built with Recharts.

Tech Stack: FastAPI • PostgreSQL / pgvector • Next.js 14 • TypeScript • RAG Pipeline

• Developed a hybrid search engine combining lexical (keyword) and vector (semantic) search over code repositories, pull requests, issues, and documentation.
• Leveraged pgvector for high-dimensional code embeddings with multi-provider LLM routing and automatic fallback mechanisms.

Tech Stack: Solidity • Foundry • Hardhat • ethers.js • EIP-712 • React

• DeFiFlow: Simulated liquidity pool and staking mechanics with gas-optimized Solidity smart contracts tested in Foundry. Implemented EIP-712 signed gasless transactions for delegated execution.
• AutoCrowd: Built for ETHGlobal ETHOnline. Decentralized crowdfunding protocol featuring automated milestone verification and smart contract escrow mechanisms.

• Founded and spearhead the AI Technology Club at Sir MVIT, conducting technical workshops, live hackathons, and collaborative coding intensives.
• Utilized the student community as a production-testing sandbox to iterate on and validate an AI Exam Proctoring System.

• ISC2 – Certified in Cybersecurity (CC): IAM, network defense, incident response, risk management.
• Cyfrin Updraft – Blockchain Fundamentals: Smart contract security, Solidity auditing, gas optimization.
• Certisured – Automation with n8n: Enterprise workflow automation & API integration.
• Cybrary – Cyber Kill Chain: Threat intelligence, attack surface reduction, mitigation strategies.

☁️ AWS Builder ID Handle: @sairaj18 Building serverless LLM pipelines, Cloud Run deployments, and cloud native architectures.	🌐 Google Developer Profile ID: 111445716260712814778 Building with Google Vertex AI, Gemini 2.0 APIs, Google Cloud Platform & Android SDKs.
🟢 NVIDIA Developer Card Status: Verified Member GPU acceleration, CUDA parallel computing, AI inference & deep learning frameworks.	🔴 AMD Developer Community Handle: @sairaj_n AMD ROCm ecosystem, hardware acceleration, heterogeneous computing & AI optimization.

# Initiating contact protocol...
$ curl -X POST https://api.sairaj.dev/v1/contact \
  -H "Authorization: Bearer OPEN_FOR_INTERNSHIPS" \
  -d '{
    "portfolio": "https://sairajmn.vercel.app",
    "x_twitter": "https://x.com/sairaj_18",
    "email": "sairaj.nadaf@gmail.com",
    "linkedin": "https://www.linkedin.com/in/sairaj-m-n/",
    "aws_builder": "https://builder.aws.com/community/@sairaj18",
    "google_developer": "https://me.developers.google.com/u/111445716260712814778",
    "nvidia_developer": "https://developer.nvidia.com/membership-card/RbQpqGtFyZFiSZEhL7XwaxSjCEh2fgl6V5y2Qvx5bXlqkCzA0fV4IDNqaZnlAytTgOWPtwj_L1U34aLQEaPHgOP_3TpTBtM=",
    "amd_developer": "https://devcommunity.amd.com/u/sairaj_n"
  }'

"Everything is an event stream until proven stateful."