Skip to content
View stym01's full-sized avatar
🎯
Focusing
🎯
Focusing
41 followers · 51 following

Achievements

Achievement: Pull SharkAchievement: Starstruck

Achievements

Achievement: Pull SharkAchievement: Starstruck

Highlights

  • Pro

Block or report stym01

Block user

Prevent this user from interacting with your repositories and sending you notifications. Learn more about blocking users.

You must be logged in to block users.

Content in all repositories owned by your account will be closed.
Maximum 250 characters. Please don’t include any personal information such as legal names or email addresses. Markdown is supported. This note will only be visible to you.
Report abuse

Contact GitHub support about this user’s behavior. Learn more about reporting abuse.

Report abuse
stym01/README.md

Hi 👋, I'm Satyam Kesharwani

👨‍💻 About Me

const satyam = {
  education: "Computer Engineering @ NIT Kurukshetra",

  interests: [
    "Systems Engineering",
    "Distributed Systems",
    "Algorithms",
    "Backend Architecture"
  ],

  currentlyLearning: [
    "Database Internals",
    "Distributed Systems Design",
    "Kubernetes",
    "Large Scale Infrastructure"
  ]

};

🚀 Quick Facts

  • 🎓 Final year Computer Engineering student at NIT Kurukshetra.
  • 💻 Building high-performance systems from scratch — C++ databases, custom allocators, distributed caching, and ML pipelines.
  • ⚔️ LeetCode Knight (Max Rating 1908).
  • 📜 Published researcher (ICDAM 2025, NE-IECCE 2026).
  • 🏫 Selected for Amazon ML Summer School 2025.
  • ⚡ Fun fact: If you ever feel useless, think about the guy who writes Terms & Conditions.
  • 📫 Reach me at: stym4193@gmail.com | LinkedIn | Portfolio

💻 Tech Stack

Languages:
C++ C Python JavaScript

Backend & Systems:
NodeJS Express.js Linux Docker AWS

Databases:
PostgreSQL MySQL Redis MongoDB

Machine Learning & Data:
TensorFlow Keras scikit-learn Pandas

🔥 Featured Projects

  • AtomicKV: Distributed Key-Value Database
    Architected a production-grade, distributed key-value database from scratch using C++17 and Linux epoll. Features a tiered storage engine (LRU Cache + Bloom Filter + B-Tree), a Consistent Hash Ring, asynchronous replication, and a masterless Gossip Protocol for high availability. Sustained 10,000+ req/sec with ~16 ms average latency under 200 concurrent clients.

  • Custom Allocators & Order Matching Engine
    Four memory allocators (Linear, Stack, Pool, Free-List) written from scratch in C++17 and used in a limit-order-book matching engine that never calls new/malloc on the matching path. Over 1M alloc/free operations, the pool allocator is ~2.5x and the linear allocator ~15x faster than glibc new/delete.

  • Shortify: Multi-Layer Caching URL Shortener
    Scalable URL shortening service with Node.js, PostgreSQL, and Redis. Engineered L1/L2 caching achieving 7.12ms latencies and a custom Token-Bucket rate limiter.

  • RansomDroid: Android Ransomware Detection
    Pioneering Vision Transformer (ViT) approach for detecting Android ransomware. Achieved 99.78% accuracy on 4,280 APKs evaluated via CuckooDroid sandbox dynamic analysis. (Published at ICDAM 2025, Springer)

  • SOC & SOH Estimation (TinyML)
    Offline diagnostic system on ESP32 using a novel Residual-Physics Neural Network (RPNN) with INT8 quantization in C++. Features real-time 'Virtual Cranking' algorithms. (Published at NE-IECCE 2026, IEEE)

🏆 Key Achievements & Publications

  • Amazon ML Summer School 2025: Selected for Amazon India's machine learning program.
  • Amazon ML Challenge 2026: Rank 313 of 26,636 teams.
  • Amazon ML Challenge 2025: Rank 2,313 of 19,556 teams.
  • LeetCode Knight: Max Rating 1908 (top 5%).
  • LeetCode Weekly Contest 497: Global Rank 412 of 35,968.
  • Publications:
    • "From Behavior to Pixels: A Vision Transformer Approach for Android Ransomware Detection" - ICDAM 2025 (Springer LNNS)
    • "SOC and SOH Estimation of Lead-Acid Battery using IoT and Residual-Physics Neural Network" - NE-IECCE 2026 (IEEE)

📈 GitHub Insights

Check out a snapshot of my coding adventures!

📊 GitHub Stats

Satyam's GitHub statistics showing commits, PRs, and contributions

🔥 Contribution Streak

GitHub contribution streak statistics

Profile view counter

Pinned Loading

  1. AtomicKV AtomicKV Public

    Distributed key-value database in C++17: single-threaded epoll server, LRU cache + Bloom filter + on-disk B-Tree, consistent hashing, gossip, Lamport clocks, read repair. 10,000+ req/s.

    C++ 2

  2. Custom-Allocator-HFT-Engine Custom-Allocator-HFT-Engine Public

    Linear, stack, pool and free-list memory allocators written from scratch in C++17, used by a price-time-priority limit order book. Pool ~2.5x and linear ~15x faster than glibc new/delete.

    C++ 2

  3. Distributed-url-shortener-with-multi-layer-caching-and-rate-limiting Distributed-url-shortener-with-multi-layer-caching-and-rate-limiting Public

    URL shortener in Node.js: in-process LRU + Redis caches over PostgreSQL (7.12 ms avg redirects), Redis Bloom filter, stampede locks, Snowflake IDs in Base62, token-bucket rate limiting. Docker on A…

    JavaScript 1

  4. Android-Ransomware-Detection-Using-Deep-Learning-ViT_CNN Android-Ransomware-Detection-Using-Deep-Learning-ViT_CNN Public

    Android ransomware detection with a Vision Transformer on CuckooDroid behaviour reports of 4,280 apps: 99.78% accuracy (ICDAM 2025, Springer). CNN and Random Forest baselines.

    Jupyter Notebook 1

  5. Iot-Project Iot-Project Public

    Offline TinyML battery monitor on an ESP32: physics-informed neural network for lead-acid SOC/SOH (R² 99.70%, IEEE NE-IECCE 2026), INT8 via Edge Impulse EON, Virtual Cranking alerts.

    C++ 1 2

  6. Next_word_prediction_using_LSTM Next_word_prediction_using_LSTM Public

    Developed a neural language model using LSTM to predict the next word in a sentence, trained on a custom corpus

    Jupyter Notebook