I am an Artificial Intelligence Researcher and Engineer focused on the intersection of Evolutionary Computation and Large Language Models. My work bridges the gap between theoretical AI research and high-performance production systems.
Currently, I am developing frameworks that use Genetic Algorithms to optimize LLM reasoning patterns, moving beyond static prompting to dynamic, evolved strategies. I also specialize in building robust, asynchronous AI infrastructure that ensures reliability and speed.
- 🔭 Current Research: Evolutionary Pedagogical Topologies (EPT) — Evolving optimal teaching strategies.
- 🎯 Focus: Neuro-symbolic AI, Agentic Workflows, and High-Throughput System Design.
- 🏆 Achievements: Winner of Vision 101 Hackathon.
A framework for evolving optimal AI teaching strategies using genetic algorithms. EPT separates the reasoning structure (Genotype) from text generation (Phenotype) to discover tutoring approaches that outperform standard Chain-of-Thought.
| Feature | Details |
|---|---|
| 🚀 The Innovation | Solves LLM "mode collapse" by evolving the structure of interactions (Diagnose → Scaffold → Hint) rather than just the content. |
| 📊 The Results | Achieved a +64.7% improvement over Chain-of-Thought baselines by discovering novel hybrid teaching strategies. |
| ⚡ System Design | Built on a fully asynchronous architecture (AsyncIO) with multi-provider fallbacks (Groq → Cerebras → OpenRouter), reducing generation time by 400%. |
Tech Stack: Python, Genetic Algorithms, AsyncIO, Hydra, Multi-Agent Orchestration.
Research & Algorithms
AI Infrastructure & Frameworks
Backend & Data
| Project | Description | Tech Stack |
|---|---|---|
| Doctor Deepseek | Fine-tuned DeepSeek-R1-Llama-8B on a medical Chain of Thought dataset. Utilized Unsloth for 2x faster training and Weights & Biases for metric logging. | Unsloth LoRA Hugging Face |
| Med Perplexity | AI-powered clinical decision support system featuring Multi-Agent architecture. Integrates PubMed API for real-time research and uses Gemini for reasoning. | LangChain FastAPI Gemini |
| Git Summarizer | An AI-powered CLI tool that generates commit messages, progress reports, and change summaries. Integrated with Slack Webhooks for automated reporting. | Python Click Rich OpenRouter |
| Local RAG Chatbot | Privacy-focused local chatbot that processes documents offline without external APIs. Uses Ollama for inference and ChromaDB for vector storage. | Ollama ChromaDB Nomic Embed |
🤖 Automation Intern | Auto Machina Remote | July 2025 – Nov 2025
- Developed comprehensive PDF parsing toolkits using Fuzzy Logic and OCR.
- Implemented cosine similarity algorithms for page-to-page image comparison.
🧠 LLM Development Intern | Viswam AI Remote | May 2025 – June 2025
- Contributed to the world’s first Telugu LLM, collaborating with Meta and the Telangana Government.
- Built scalable data collection platforms and fine-tuned Whisper models for audio transcription.