Raghu Talla

Computer Engineer @ Texas A&M

Software Engineering
System Architecture
Hardware Design

About Me

I chose to study computer engineering because I like to code but I'm also fascinated with how our hardware lets us do whatever we can imagine. In my free time, I like to watch YouTube videos explaining how our technologies work and other random stuff. Of note is Wintergatan , documenting his journey of making a musical Marble Machine. I find the combination of engineering challenges with music to be pretty cool.

I also like to play my violin, play badminton, and work out.

Experience

GenAI Technical Advisor Intern

Scale AI | November 2024 - December 2025

  • Enhanced reasoning and code-generation of LLMs deployed by leading tech companies, improving accuracy and efficiency on competitive programming tasks
  • Designed and executed RLHF workflows (reward model calibration, ranking tasks) to refine LLM outputs for reasoning, research, and multi-hop problems
  • Benchmarked state-of-the-art models on long-context and deep-research prompts, delivering insights that shaped enterprise deployment strategies

SWE + Data Engineering Intern

Aris Insights | May 2025 - August 2025

  • Engineered ClickHouse BI analytics platform with error prevention for 7+ error codes, reducing query failures by 95% and delivering production-ready reliability to fintech clients
  • Implemented a progressive complexity framework with automatic fallback strategies, achieving 100% uptime in customer deployments
  • Built data generation framework simulating transaction patterns and lifecycle events with Python, NumPy, and statistical modeling; optimized batch processing with PostgreSQL, improving processing efficiency by 65%
  • Collaborated in an Agile workflow using Linear for sprint planning, issue tracking, and feature delivery

Instructor

iCode | June 2024 - August 2024

  • Taught game development, web development, and UI/UX design to middle and high school students using Construct 3, Replit, and Figma
  • Designed and delivered hands-on lessons that guided students from foundational coding concepts to completed projects
  • Mentored students in debugging, problem-solving, and teamwork while developing strong communication and leadership skills

Skills

Languages

  • Verilog
  • OpenCL
  • C++
  • Assembly (RISC-V, ARM)
  • Bash/Linux
  • Java
  • MATLAB
  • Python
  • Typescript
  • R
  • SQL
  • HTML/CSS

Frameworks & Libraries

  • Node.js
  • React.js
  • Next.js
  • Flask

Developer Tools

  • Git
  • Vivado
  • SPICE
  • OpenStack
  • Docker
  • PostgreSQL
  • ClickHouse

Featured Projects

Secure UART Bootloader & Bare-Metal Application (RISC-V)

July 2025 - Present

  • Built a secure UART bootloader with flash programming, SHA-256/Ed25519 verification, and rollback protection
  • Developed a bare-metal application with custom linker/startup code and interrupt vectors
  • Created a Python host flasher (pyserial, COBS, CRC32) and debugged via JTAG and UART

Embedded & Systems Programming Projects

August 2025 - Present

  • Built multithreaded C++ apps using POSIX threads, mutexes, and condition variables for real-time reliability
  • Developed a Unix shell with process creation, I/O redirection, and TCP/UDP sockets; deployed on Ubuntu VM with OpenStack and GDB

GPU-Accelerated Conway's Game of Life

May 2025

  • Developed a C++/OpenCL simulator for John Conway's Game of Life with parallelized cell updates
  • Optimized 1-byte-per-cell kernel, delivering ~1.5-2.5x faster performance vs. bit-packed baseline on 2k-4k maps

IoT Appliance-Localized Fire Suppression System

June 2023

  • Programmed ESP8266 microcontroller to integrate IR flame sensor, relays, and a linear actuator for automated, appliance-level fire suppression
  • Implemented IoT connectivity with real-time mobile alerts for immediate user response

RAG LLM Dementia Reminiscence Therapy ChatBot

TAMUhack 2025

  • Developed RAG-based chatbot to generate personalized conversations for maintaining cognitive function in patients by incorporating family-provided memories
  • Leveraged Next.js, AWS S3 and Bedrock, and Amazon's Titan Text Embeddings V2 model

OpenLedger: FinTech Transparency Platform

October 2025

  • Developed a full-stack Next.js + Supabase platform that scans fintech codebases and auto-generates verifiable privacy policies
  • Integrated Gemini multi-agent workflows and telemetry for versioned disclosures
  • Designed modular API endpoints and responsive UI for live policy generation and transparency reporting