MS Computer Engineering @ ASU

Ram Harikrishnan

AI Systems Engineer · Applied ML Engineer · Full-Stack Developer

Systems-focused Software Engineer with 4+ years of experience in embedded systems, C/C++ programming, and hardware-software integration. Building production-grade ML systems and intelligent applications. Winner of AMD's Synthetic Data AI Agents Challenge. Specializing in reinforcement learning, GraphRAG architectures, and scalable AI infrastructure.

View Projects Get in Touch

Professional Experience

4+ years building scalable systems and intelligent applications

EXPERIENCE

−

□

May 2024 - Present

Full Stack Developer

Arizona Water Project - Arizona State University

Tempe, AZ

Developed and deployed statewide FastAPI/React system on AWS (EC2, S3) with 50,000+ user interactions
Built Arizona WaterBot, a multilingual RAG chatbot using LangChain, OpenAI, and Neo4j with auto-content ingestion
Performed extensive debugging and root cause analysis using structured methodologies to improve system reliability
Optimized PostgreSQL queries and backend logic to reduce latency and stabilize performance under peak load
Implemented CI/CD pipelines using GitHub Actions, improving release quality and deployment consistency
Created Python and Bash automation scripts to streamline data processing and deployment workflows

EXPERIENCE

−

□

May 2022 - October 2023

Full Stack Developer

PIT Solutions Pvt. Ltd.

Trivandrum, India

Engineered Python (Django) microservices for large-scale AI and data processing workflows
Led migration from legacy .NET monolith to microservices architecture using React and Spring Boot
Built automation scripts (Python, Bash, PowerShell) to streamline deployment and improve development velocity
Performed CPU/memory profiling and structured debugging to identify bottlenecks and resolve performance issues
Designed secure REST APIs with OAuth2.0/JWT authentication connecting Oracle and MS SQL databases
Automated CI/CD pipelines using Jenkins for legacy and modern systems; developed test suites using Cypress and Jest

EXPERIENCE

−

□

December 2020 - April 2022

Software Engineer & Co-founder

MakerLabs Edutech Pvt. Ltd.

Kochi, India

Co-founded and architected IoT-enabled STEM learning platform with C/C++ firmware and Python backend infrastructure
Designed and developed ARM Cortex-M based smart educational toys integrating sensors, actuators, and wireless communication modules
Implemented hardware-software integration layer for ARM-based systems, bridging low-level firmware with high-level application logic
Developed embedded firmware for Arduino and ESP32 microcontrollers using C/C++, implementing interrupt-driven programming
Built real-time communication protocols (MQTT/WebSockets) enabling seamless embedded device-to-cloud connectivity

Projects & Work

Showcasing impactful ML engineering projects and technical achievements

FEATURED PROJECT

−

□

🏆 AMD/META HACKATHON WINNER

Synthetic Data Generation Pipeline for LLM Reasoning

Built a multi-stage synthetic data generation pipeline for the Meta AI/AMD/PyTorch Synthetic Data AI Agents Challenge. Designed a relationship-aware system that generates high-quality question-answer pairs for training LLMs on reasoning tasks involving character relationships and seating arrangements. Implemented Guided Reinforcement Policy Optimization (GRPO) with custom reward functions, structured JSON output enforcement, and identified critical LLM vulnerabilities like alphabetical sorting bias. Fine-tuned multiple models (Llama 3 70B, Qwen3-Next 80B MoE, Phi-4) using LoRA, SFT, and GRPO techniques.

📖 Read Technical Deep Dive →

16K+

Lines of Code

1st

Place Winner

95%+

Test Coverage

LLM Models

Python PyTorch GRPO LoRA Fine-tuning Llama 3 Qwen3 Phi-4 Structured Output Reinforcement Learning

PROJECT

−

□

Startup Intelligence Analysis App

Production-ready GraphRAG application for startup intelligence analysis. Built with FastAPI backend, Neo4j graph database, and OpenAI GPT-4o for intelligent query processing. Features 40+ API endpoints, Redis caching, Prometheus monitoring, and comprehensive test coverage. Deployed on Vercel with Docker containerization and CI/CD pipelines.

GraphRAG FastAPI Neo4j OpenAI GPT-4o Redis Prometheus Docker TypeScript

GitHub → Live Demo →

PROJECT

−

□

Wildfire Control RL Agent

Reached Final Four with a reinforcement learning system using Llama 3.2 with LoRA fine-tuning. Trained on 16,000+ demonstrations using GRPO and behavior cloning on AMD MI100 GPU. Deployed to Hugging Face Spaces with optimized inference.

Reinforcement Learning Llama 3.2 LoRA GRPO Hugging Face

GitHub →

PROJECT

−

□

Graph Processing at Scale

Built large-scale graph data processing pipeline using Neo4j, Terraform, AWS, Docker, Kubernetes, and Kafka. Processed NYC Yellow Cab Trip dataset, transforming raw records into structured graph nodes and relationships. Implemented PageRank and BFS algorithms using Neo4j Graph Data Science to analyze node importance and connectivity.

Neo4j Kubernetes Kafka AWS Terraform Graph Algorithms

PROJECT

−

□

AI-Generated Image Classification

Developed deep learning model to classify AI-generated and real images using ResNet50v2 and EfficientNetV2. Preprocessed datasets (CFAKE & ArtiFact) with data augmentation techniques. Performed hyperparameter tuning and evaluated using AUC, accuracy, precision, recall, and F1-score metrics.

Deep Learning CNNs ResNet Transfer Learning PyTorch

PROJECT

−

□

Embedded Systems & IoT Integration

Developed C/C++ firmware for ARM Cortex-M and AVR microcontrollers for IoT sensor networks and robotics applications. Implemented interrupt-driven programming and low-level peripheral drivers (UART, SPI, I2C) for sensor integration. Applied systematic debugging methodologies using JTAG debuggers and serial monitors.

C/C++ ARM Cortex-M Embedded Systems IoT Firmware

Technical Expertise

Full-stack ML engineering and software development

SKILLS

−

□

🤖 Machine Learning & AI

PyTorch & TensorFlow

Reinforcement Learning

LLM Fine-tuning (LoRA, GRPO)

RAG & GraphRAG Architectures

LangChain & LlamaIndex

CNNs & Deep Learning

Model Optimization

SKILLS

−

□

⚙️ Backend & Infrastructure

Python (FastAPI, Django, Flask)

Node.js & Express

Spring Boot

Docker & Kubernetes

CI/CD Pipelines (Jenkins, GitHub Actions)

Database Design (SQL/NoSQL)

Neo4j Graph Databases

Terraform

SKILLS

−

□

🎨 Frontend Development

React & TypeScript

Vue.js & Angular

Next.js

JavaScript (ES6+)

HTML5 & CSS3

Electron.js

Responsive Design

SKILLS

−

□

💻 Systems & Embedded

C/C++ Programming

ARM Cortex-M Microcontrollers

Embedded Systems & Firmware

RTOS Concepts

Hardware-Software Integration

Interrupt-Driven Programming

PC/CPU Architecture

SKILLS

−

□

☁️ Cloud & DevOps

AWS (EC2, S3, Cloud Services)

Azure

Docker & Kubernetes

Jenkins & GitHub Actions

Terraform

Linux/Unix (Ubuntu, CentOS)

Shell Scripting (Bash, PowerShell)

SKILLS

−

□

🔧 Tools & Technologies

Git, GitHub, Perforce

PostgreSQL, MySQL, MongoDB

Oracle, MS SQL, Redis, Firebase

OAuth2, JWT, SSL/TLS

WebSockets, MQTT

Testing (Cypress, Jest, Mocha)

JTAG Debuggers

Education

Academic background and coursework

EDUCATION

−

□

Master of Science in Computer Engineering

Arizona State University

Tempe, AZ

GPA: 3.3/4.0

Expected: December 2025

Relevant Coursework:

Operating Systems Computer Architecture Distributed Systems Data Processing at Scale Statistical Machine Learning

EDUCATION

−

□

Bachelor of Technology

PRIST University

Thanjavur, India

CGPA: 7.2/10

March 2020

Major:

Electronics & Communication Engineering

Achievements

Recognition and contributions

ACHIEVEMENT

−

□

2025

1st Place — Meta AI / AMD / PyTorch Synthetic Data Hackathon

Built novel multi-stage synthetic data pipeline for complex relational reasoning tasks using Python and PyTorch. Implemented Guided Reinforcement Policy Optimization (GRPO) with custom JSON rewards for logical accuracy. Fine-tuned Llama 3, Qwen 3, and Phi-4 models using LoRA, creating separate Q/A model pairs. Identified and mitigated major LLM reasoning flaw (alphabetical sorting bias), improving logical consistency.

📖 Read Technical Deep Dive →

ACHIEVEMENT

−

□

2025

Open Source Contributions — Meta PyTorch OpenEnv

Developed wildfire simulation environment enabling reinforcement learning research and agent training. Collaborated with Meta/PyTorch teams to enhance simulation capabilities and documentation. Wrote comprehensive technical documentation and unit tests, improving adoption and development workflows.

CONTACT

−

□

Let's Connect

Currently seeking software engineering co-op opportunities for Spring/Summer 2026 or Summer/Fall 2026. Open to discussing ML engineering, full-stack development, embedded systems, and innovative AI projects.

Work Authorization: Eligible for CPT/OPT (F-1 visa status)
Location Preference: Open to Austin, TX; Boxborough, MA; Fort Collins, CO; Longmont, CO; Fishkill, NY; Rochester, NY