Tanay Tammineni TammineniTanay

About Me

#!/usr/bin/env python3
# tanay_tammineni.py

class AIEngineer:

    def __init__(self):
        self.name       = "Tanay Tammineni"
        self.role       = "AI Systems Engineer"
        self.education  = "MS CS @ SEMO · 3.9 GPA"
        self.location   = "Irving, TX · Open to Relocate"
        self.status     = "🟢 Open to Opportunities"

    @property
    def expertise(self):
        return {
            "LLM"  : ["QLoRA", "DeepSpeed ZeRO-3",
                      "vLLM", "Flash Attention 2"],
            "RAG"  : ["Qdrant", "Elasticsearch",
                      "Neo4j", "LangGraph", "CRAG"],
            "Data" : ["PySpark", "Databricks",
                      "SQL", "Pandas", "RAGAS"],
            "Cloud": ["AWS", "Docker",
                      "Terraform", "CI/CD"],
        }

    @property
    def achievements(self):
        return {
            "memory_reduction"  : "41.2% per-GPU",
            "throughput_gain"   : "3.8x on Llama 3 8B",
            "faithfulness_gain" : "+23.7% via RRF",
            "publications"      : 2,
            "gpa"               : 3.9,
        }

    def say_hi(self):
        print("I don't just build AI. I ship it.")


me = AIEngineer()
me.say_hi()

⚡ Key Metrics

🔧 Currently Building

🤖 JobAgent

Zero-cost job application pipeline. Local Ollama · llama3.1:8b · SQLite · LaTeX. No API calls. No cost.

Ollama SQLite Pandas LaTeX

🎙️ LiveWire AI Co-Pilot

Chrome MV3 extension · Tab + mic capture · Whisper STT · Evidence packs.

WebSocket FastAPI Whisper Chrome MV3

📄 UniLLMOps Framework

Unified LLM production framework — fine-tuning to serving. Zenodo · Targeting arXiv cs.AI.

LLM RAG CRAG RAGAS vLLM

🛠️ Tech Stack

Core AI/ML

Infrastructure & Cloud

Databases & Search

Dev Tools

🚀 Flagship Projects

⚡ Distributed LLM Fine-Tuning Pipeline

Metric	Result
💾 Per-GPU memory	−41.2% via ZeRO-3
⚡ Throughput	3.8x on Llama 3 8B
🔬 Techniques	DPO · TIES · DARE · SLERP
📦 Infra	Prometheus · Grafana · CI/CD