CT. Chanakya Thotakura

Security Engineer · Detection as Code · AWS + Cloudflare

Chanakya Thotakura

Security & Full-Stack Engineer

I am Chanakya Thotakura, a software engineer working where security, AI and product meet. I build endpoint detection tooling, encrypted consumer applications, retrieval-backed copilots and the cloud infrastructure underneath them, in Python, TypeScript, React Native and Swift across AWS, GCP and Cloudflare.

Selected work Résumé (PDF) Interactive portfolio

Maryville, Missouri · open to security engineering roles · M.S. Applied Computer Science, Northwest Missouri State University

About

The short version, for anyone who landed here from a search for my name.

Chanakya Thotakura is a security-minded full-stack engineer based in Maryville, Missouri, currently completing an M.S. in Applied Computer Science at Northwest Missouri State University. My work centres on detection engineering — writing and version-controlling Sigma-style rules for phishing, insider fraud, impossible travel and endpoint signals — and on control assurance for EDR platforms including CrowdStrike, Microsoft Defender and SentinelOne.

Alongside that I ship product. ManaSplit is an iOS application on the App Store; the Northwest Student Planner is a retrieval-augmented copilot bounded to cited records; Visu-Net makes neural network training a thing you can watch happen in a browser tab. The common thread is systems that explain themselves: evidence-bounded triage, citation-bound answers, and encryption that holds up when you look closely at the key management.

Previously a Machine Learning Intern at AIMERS and a Cloud Engineering Intern working on BigQuery data-warehouse optimisation. I also run self-hosted infrastructure and have spent real time terminating Cat6 and bringing up enterprise access points, which is where most of my instincts about reliable systems actually came from.

Selected work

Six of thirteen shipped builds. Full case studies, with tradeoffs and outcomes, live on the interactive portfolio.

ControlForge Security

Security · Creator & Security Engineer · 2025–present

A Python platform that verifies endpoint control health across CrowdStrike, Defender and SentinelOne, then evaluates normalised events against version-controlled, Sigma-style YAML detections covering phishing, insider fraud and edge signals. The CLI and FastAPI service return stable alert IDs, matched evidence and investigation-ready reasons, so every alert can be traced back to the rule and the data that produced it. Verified-domain dark web and infostealer monitoring runs on AWS and Cloudflare.

  • Python
  • FastAPI
  • Sigma
  • YARA-comparable
  • AWS S3
  • Docker

GitHub repository Case study

ManaSplit — group expenses and shared conversations

Product · Creator & Full-Stack Engineer · App Store · 2024–present

An iOS-first app where the expense, the conversation and the call live in one place. Eleven split methods (equal, shares, percentage, custom, itemised receipts, time-based and a gamified roulette) run through a typed state machine in strict TypeScript. Message history stays on device — Firebase RTDB is an ephemeral transit queue, never the source of truth. Native calling comes from RTDB signalling plus LiveKit media and CallKit/PushKit, so calls ring properly even from a terminated app. An opt-in Security Center adds HIBP verified-domain breach checks, Flare dark web signals and Signal E2EE per linked device.

  • Expo 55
  • React Native
  • TypeScript
  • Firebase
  • LiveKit
  • Cloud KMS

Case study

Northwest Student Planner — secure RAG copilot

Product · Developer, 5-person graduate team · 2026

An offline-capable planner whose copilot is bound to cited records or abstains entirely. React/Vite web and Expo mobile clients share one FastAPI surface on GCP Cloud Run; Firebase ID tokens gate it and the user ID is derived only from the verified token. Offline storage is encrypted at the edge — non-extractable AES-GCM keys in IndexedDB on web, XChaCha20-Poly1305 via SecureStore on mobile — and Firestore payloads use per-user AES-256-GCM keys wrapped by Cloud KMS. Retrieval runs UID-prefiltered KNN over Vertex embeddings, and planner mutations stay typed, previewed proposals that apply only on explicit confirmation.

  • React
  • FastAPI
  • Vertex AI
  • Cloud Run
  • Firestore + KMS
  • MCP

Live GitHub Case study

Visu-Net — interactive neural network visualiser

AI · Creator · 2025

A neural network you can pick up and turn over. Design the architecture by drag-and-drop, train it in the browser with brain.js, and watch weights, activations, metrics and the loss landscape update live — all tunable mid-training, with no backend and no server round trip. The same pattern ports directly to edge fraud and risk-scoring interfaces.

  • React
  • TypeScript
  • brain.js
  • Recharts
  • MUI

Live demo GitHub Case study

TeluguChatBot — real-time voice assistant

AI · Solo developer, extended from the AIMERS internship · 2024–present

A Telugu-language voice assistant that routes spoken queries either to live data sources or to a retrieval knowledge engine, with edge inference and phishing-resistant intent allowlists so the model can only act inside a vetted command surface.

  • Python
  • RAG
  • Speech
  • Edge inference

GitHub Case study

C++ to Java converter

AI · Solo developer · 2025

Import a single C++ file or an entire repository through GitHub OAuth, queue repo-wide conversion jobs in Redis, and convert with Gemini while preserving directory structure. Converted code is reviewable in-browser, executable in sandboxed Docker containers, and pushable straight to a new branch — with a rule-based Python fallback for when the model is the wrong tool.

  • Python
  • Gemini
  • Redis
  • Docker
  • GitHub OAuth

GitHub Case study

Experience & education

Where the work above came from.

  • Fall 2026 Developer — Northwest Student Planner (GDP 2) Graduate Directed Project II, 5-person team. Deepening schedule optimisation and the companion mobile app.
  • Summer 2026 Developer — Northwest Student Planner (GDP 1) Shipped v1.0–1.1: widget dashboard, multi-view calendar with drag-and-drop rescheduling, deadline-based priority escalation and overloaded-day detection.
  • Spring 2026 iOS Developer — collaborative budgeting app Four-person team, M.S. iOS course.
  • 2025 – 2027 M.S., Applied Computer Science — Northwest Missouri State University Database systems, software engineering, data analytics and mobile development.
  • Jun – Aug 2024 Machine Learning Intern — AIMERS Speech and retrieval work that became TeluguChatBot.
  • 2024 Cloud Engineering Intern Six-week enterprise internship on GCP BigQuery data-warehouse optimisation.
  • 2024 Web Lead & Organiser — AFOSEC National College Fest Built and ran the registration platform for a national student festival.
  • Ongoing Self-hosted infrastructure & networking Running my own hardware: Docker behind Nginx, structured cabling, enterprise access points.

Skills

What I reach for, roughly in order of how often I reach for it.

Security
Detection engineering, Sigma-style rules, endpoint control assurance (CrowdStrike, Defender, SentinelOne), breach and dark-web monitoring, phishing and URL analysis, IAM, Signal E2EE
AI
RAG, Model Context Protocol, Gemini and Vertex AI, embeddings and vector search, prompt-injection defence, structured outputs, evaluation
Cloud & backend
AWS, GCP, Firebase, Cloudflare, Docker, FastAPI, REST, Cloud KMS, Secret Manager, SQLite
Languages
Python, TypeScript and JavaScript, Swift, Kotlin, SQL
Frameworks
React, React Native and Expo, Vite, GitHub Actions

Contact

Open to security engineering roles. The fastest way to reach me is email.

chanikya1@icloud.com LinkedIn GitHub Résumé (PDF)