Systems Engineer & AI Researcher | Specializing in Agentic Workflows, DevOps, & Financial Intelligence

Mastering the intersection of Applied Mathematics and Systems Engineering to build autonomous, high-performance environments.

I am currently deep-diving into the mathematical and programmatic foundations required to scale AI and Finance. My work bridges the gap between raw data and actionable intelligence through optimized backend architecture and agentic automation.

• 🔭 Current Focus: Engineering the core fundamentals of Math and Programming for AI Agentic Development, DevOps automation, and Quantitative Financial Analysis.
• 🧠 Philosophy: "First Principles Engineering"—understanding the underlying math (Linear Algebra, Calculus, Statistics) to build more efficient AI and financial models.
• 🏡 Infrastructure: Operating a production-grade Homelab to stress-test DevOps workflows and host private, local-first AI agents.

Solving data fragmentation and latency in high-frequency financial analysis.

• The Problem: Processing massive, multi-exchange crypto data for technical analysis is often bottlenecked by database I/O and unoptimized calculation logic.
• The Solution: A high-performance pipeline leveraging TimescaleDB for time-series optimization and Celery for distributed processing. It automates ingestion and calculates 50+ indicators across multiple timeframes using vectorized math.
• Tech: Python 3.12, TimescaleDB (PostgreSQL 17), Redis, CCXT, Pandas-TA, Docker.

Solving the privacy and latency barriers in assistive technology.

• The Problem: Most AI accessibility tools rely on cloud APIs, creating privacy risks and unacceptable delays in real-world environments.
• The Solution: A local-first AI suite that uses On-Device Inference to perform real-time Computer Vision (object/gesture recognition). It demonstrates the power of running "Agentic" logic at the edge without internet dependency.
• Tech: Next.js, MediaPipe, TensorFlow Lite, Tailwind CSS.

Solving institutional friction through programmatic optimization.

• The Problem: Manual resource allocation (like seating arrangements) is a combinatorial problem that is tedious and error-prone for humans.
• The Solution: SeatGenie, a tool that uses algorithmic optimization to generate complex seating plans in seconds, ensuring privacy through a local-first architecture.
• Tech: Next.js, React, TailwindCSS, TypeScript.

• AI & Agentic Dev: LangChain, Agentic Workflows, MediaPipe, On-Device Inference, Applied Math for AI.
• DevOps & Infra: Docker/Compose, CI/CD, Nginx Proxy Manager, Linux SysAdmin, Prometheus/Grafana.
• Financial Analysis: Time-series Data (TimescaleDB), Vectorized Operations (Pandas/NumPy), Quantitative Indicators.
• Data Mastery: PostgreSQL (Advanced/Replication), Redis (Streams/Caching), RabbitMQ, Data Modeling.

• 📧 xotembotzop@gmail.com | 🌍 surjyadipsen.in
• 💼 Freelancer Profile | 📦 GitHub Repos

"Automate the mundane, create the extraordinary."
I build my own tools and infrastructure from scratch to ensure I understand every layer of the stack, from the hardware to the high-level AI logic.