Hello, I'm

Harsh Maheshwari

Software Engineer & Researcher

Computer Science & Computational Mathematics student at Penn State. Building intelligent systems and pushing the boundaries of AI.

View My Work

About Me

I'm a Computer Science and Computational Mathematics student at Pennsylvania State University with a passion for software engineering, machine learning, and research. Currently maintaining a 3.8/4.0 GPA.

My experience spans from developing AI-driven platforms to conducting research on bias in large language models. I've worked with diverse technologies including React Native, AWS, Flask, and various machine learning frameworks.

As a former CTO at Marble, I led development teams using Agile methodologies and shipped products that make a positive impact. Currently, I'm working on frontier AI safety at Scale AI and building healthcare solutions at SIGNUM.

Experience

My professional journey across AI, software engineering, and research.

Jun 2025 – Present

GenAI Technical Advisor Intern

Scale AI

San Francisco, CA

  • Prompt-engineered frontier LLMs; wrote 150+ tests (reasoning, tool-use, safety, retrieval); fixes cut jailbreaks/hallucinations by 20–30%.
  • Systematically tested LLM tool-calling strategies; rewrote instructions and examples, improving model's adherence to complex instructions by 12% and decreasing tool errors by 10%.
LLMsPrompt EngineeringAI SafetyTool-Use Testing
Aug 2025 – Present

Software Engineer

SIGNUM

State College, PA

  • Designed and developed a React Native app; Migrated backend services on AWS increasing the security and efficiency by 20%.
  • Engineered and launched geofencing and wearable event processing features, increasing caregiver response time to critical events by 25% and contributing to enhanced patient safety and well-being.
React NativeAWSGeofencingWearables
April 2024 – Present

Research Intern

College of Information Science and Technology

State College, PA

  • Publishing a research paper as the lead author on the bias by LLM's on 20+ stereotypical subjects by Llama 2 7B using graphs.
  • Developed Python scripts and Jupyter notebooks to automate data collection, leveraging Flask API to retrieve model responses from various LLM models on Hugging face
PythonJupyterFlaskLLMsLlama 2
Sep 2023 – Feb 2025

Chief Technical Officer

Marble

State College, PA

  • Launched an AI-driven mental-health platform; led a Next.js/React/TypeScript web MVP with Node.js APIs and Flask microservices integrating AWS, Firebase, OpenAI, and ElevenLabs.
  • Shipped MVP in under 4 weeks (team of 4), achieved 50+ downloads in week 1, incorporated feedback from 300+ user interviews, and secured $7.5K in awards.
Node.jsNext.jsReactTypeScriptFlaskAWSFirebaseOpenAI
May 2024 – Aug 2024

Software Engineering Intern

Penn State College of Medicine

State College, PA

  • Developed a cross-platform React Native clinical-support app with Firebase synchronization and role-based access; coordinated 20+ beta tests and shipped on time.
  • Collaborated in an agile team of 5; delivered reusable UI components and analytics aligned with clinician workflows.
React NativeFirebaseAgileUI Components
Jan 2024 – Apr 2024

Software Engineering Intern

World Hope

State College, PA

  • Architected a Python RAG backend with MongoDB enabling fast retrieval and AI Q&A for caseworkers; designed for 1,000+ field users serving 100,000+ beneficiaries.
PythonRAGMongoDBLangChainAI

Education

Academic foundation in computer science and mathematics.

Pennsylvania State University

State College, PA

Aug 2022 – May 2026

B.S. in Computer Science; B.S. in Computational Mathematics

GPA: 3.8/4.0

Awards

President Walker AwardPresident Sparks Award

Relevant Coursework

Data Structures & AlgorithmsMachine LearningArtificial IntelligenceDatabase SystemsLinear AlgebraDiscrete MathematicsProbability & Statistics

Projects

A selection of projects I've built, from AI-powered tools to award-winning hackathon submissions.

Zenny.ai
2024 – Present

Predictive execution co-pilot for Slack/Teams that connects to Jira, Slack, OKR platforms, and wikis to detect misalignment and risks early. Created a working MVP within 1 month. Impact: earlier blocker detection; managers save approximately 4 hours/week on status updates.

AWSPythonLangChainVector DBAirweave API
Watch Demo
Nebula
Oct 2023

Built tumor and Alzheimer's classifiers on 300+ DICOM CTs with 3D visualization via VTK; achieved accuracies up to 94% and 80%. Won 1st place in Generative AI category at HackPSU.

FlaskHTML/CSS/JSCNNVTKNumPyTensorFlow
View Source
Signease
March 2023

Collaborated with a team of five to design and create a real-time ASL Translator app with an accuracy of 87%, delivering third-place success in a competitive Hack PSU event.

NumPyMatplotlibTensorFlowRandom ForestComputer Vision
View Source

Skills

Technologies and tools I work with regularly.

Languages

JavaPythonTypeScriptJavaScriptC/C++KotlinDartRSQLHTML/CSSBashVerilog

Frameworks & Libraries

ReactReact NativeNext.jsNode.jsFlaskTensorFlowPandasNumPy

Tools & Technologies

MongoDBFirebaseAWSAzureGitLinux

Let's Connect

I'm always open to discussing new opportunities, interesting projects, or just having a conversation about technology and AI.

Send me an email