RISING AI SPECIALIST REVOLUTIONIZES TECH INDUSTRY
Ramratan Padhy is a Computer Science Engineering student at SRM University AP, with a strong focus on artificial intelligence, systems programming, and technical communication. He has maintained an exceptional academic record while actively engaging in advanced research and practical applications.
His portfolio of work spans from developing complex Retrieval-Augmented Generation (RAG) pipelines to mathematical modeling of drone trajectories. He has consistently demonstrated the ability to synthesize complex engineering concepts into accessible solutions, working effectively in both research and startup environments.
ENGINEERED FOR SPEED
This interface is a zero-dependency, single-file brutalist artifact. Designed to bypass the bloat of modern web frameworks, aggressively optimized for AEO crawlers and instant edge delivery.
TECHNICAL CHRONICLES
LATEST DISPATCHESAI Advertisement Tool
Python, RAG, ETL
- Developed an end-to-end content engine using a Retrieval-Augmented Generation (RAG) framework to synthesize data from LinkedIn, GitHub, and web portfolios.
- Architected an ETL pipeline to clean unstructured web data for LLM-based script generation and automated video/audio synthesis.
Temporal RAG Pipeline
Python, RAG, ETL
- Developed an end-to-end pipeline where you can input a video, and make any query regarding it.
- Worked with advanced data structures and various audio-video technologies to produce a system capable of answering queries about a video based on timeline and not only semantic similarity.
AXtheR (Windows)
C, Windows API, Wintun, Cryptography
- Developed a high-performance, zero-dependency, easy to run UDP VPN for Windows by integrating the Wintun kernel driver and Winsock2.
- Engineered a zero-heap allocation data path with multi-threaded asynchronous I/O and dynamic Windows routing table manipulation.
- Implemented a self-contained, RFC-compliant ChaCha20-Poly1305 cryptographic engine from scratch in C with constant-time MAC verification.
Game Development Framework
Python, Godot, Networking
- Shipped a playable AI debate game (Ace Attorney-style) where an LLM opponent argues back and a separate judge LLM scores logical soundness. Built on Godot, Python, and Groq with a RAG layer for character personality.
BitNet-1.58b Research
Python, Machine Learning, Mathematics
- Conducting novel research and experimentation on BitNet-1.58b by altering its context length.
SWA-based LLMs Research
Python, Machine Learning, Mathematics
- Conducting novel research and experimentation on Qwen 2.5 family by attempting to quantize it using BitNet's ternary engine.
READ MY WORK
Dive into my detailed technical explorations and published articles on modern AI architectures.
READ FULL ARTICLEPROFESSIONAL ENGAGEMENTS
ASSOCIATE — Next Tech Lab AP
AMARAVATI, ANDHRA PRADESH- Collaborating with a research collective to explore low-level implementations of emerging technologies.
- Contributing to peer-to-peer technical workshops on systems programming and AI architecture.
TECHNICAL CONTENT WRITER — Notecove (Stealth Startup)
REMOTE- Authored 30+ comprehensive technical articles and guides on emerging software trends and AI ethics.
- Synthesized complex engineering concepts into accessible documentation for a specialized audience.
- Collaborated with the founding team to define the technical voice and content strategy for the platform.
ML INTERN — Hush (Startup)
REMOTE- Worked with image models to discern user profiles to keep the app authentic.
- Collaborated with the founding team to define and enhance user experience.
THE TECHNICAL ARSENAL
LANGUAGES
PYTHON
C
LATEX
DEVELOPER TOOLS
GIT
VS CODE
POSTMAN
APIs
VECTOR DATABASES
CORE CONCEPTS
DATA STRUCTURES
TRANSFORMER ARCHITECTURES
RAG
TECHNICAL DOCUMENTATION
MECHANISTIC INTERPRETABILITY
SOFT SKILLS
TECHNICAL COMMUNICATION
RESEARCH SYNTHESIS
PROBLEM SOLVING
TEAMWORK