Computer Science @ CSULB · Software / AI Engineer
Production AI agents · LLM safety & security · Full-stack & backend · Multithreading & concurrency

• 🏢 Software Engineering Intern @ ABB — building a production, ServiceNow-integrated enterprise AI agent that classifies, triages, and auto-resolves support tickets with confidence-based routing.
• 🧠 AI Engineer Intern @ Simmspace.AI — building a developer platform (Simmspace IDE + SimmSilos) on a scalable FastAPI backend, plus a security-focused AI risk layer for code submissions.
• 🔬 Undergraduate Research Assistant @ CSULB (prior) — multithreading, concurrency, and parallel performance in Java & C++.
• 📚 Instructional Student Assistant @ CSULB — helped 200+ students strengthen their Python and math foundations.
• 🎯 I like building at the intersection of AI systems, safety/security, and clean backend design — and I'm always up to learn from open-source and collaborate on impactful projects.

🎓 B.S. Computer Science, CSULB · Jan 2024 – Dec 2027

Languages

AI / ML

RAG · LLM evaluation · model training · inference pipelines · agent workflow design

AI Safety & Security

Prompt-injection scanning · PII detection · hallucination checks · confidence-based routing · audit logging · threat detection

Backend & MLOps

Systems & Tooling

Multithreading · concurrency · benchmarking · automated testing · REST APIs · CI/CD

Python · PyTorch · FastAPI · Gemini API

A closed-loop system for network-traffic anomaly detection that trains, evaluates, and reruns experiments to improve accuracy.

• Closed-loop PyTorch training that collects data, evaluates outputs, identifies failure modes, and iterates.
• Integrated the Gemini API to convert detected attack patterns into WAF security rules via confidence-based agent workflows.
• Served low-latency inference through a FastAPI endpoint using ML/NLP techniques for real-time threat identification.

Python · LangChain · ChromaDB · Gemini API · FastAPI

An agentic RAG pipeline that ingests documents and produces grounded, self-evaluated LLM answers.

• Ingests PDF, DOCX, and TXT, generates local embeddings with Sentence Transformers, and retrieves context via ChromaDB.
• Designed self-evaluation and confidence-scoring loops that rerank context, detect failure modes, and iterate for reliability.
• Exposed research workflows through a FastAPI backend with semantic indexing and retrieval-quality evaluation.

JavaScript · Web Audio API · HTML/CSS

A browser-based harmonium simulator with realistic bellows mechanics and multi-octave playback.

• Uses the Web Audio API with gain nodes and reverb for dynamic tone generation.
• Real-time keyboard and mouse controls and a responsive UI for volume, reverb, and octave selection.

Python · OOP · Data Structures

A turn-based battle game showcasing object-oriented design with modular Pokémon, Item, and Player classes, battles, an item shop, and health tracking.

Java · Data Structures · OOP

A bookstore management system using HashMaps, Linked Lists, Queues, and Stacks for efficient search, checkout, and cart operations.

🏢 Software Engineering Intern — ABB · May 2026 – Present · Cary, NC

• Developing a production-facing, ServiceNow-integrated enterprise AI agent to classify, triage, and auto-resolve tickets, targeting a 30% reduction in org-wide ticket volume.
• Designed confidence-based routing: resolves tickets only on strong knowledge-base retrieval, escalating ambiguous cases to humans.
• Built live analytics dashboards over the ServiceNow REST API and instrumented the pipeline with structured logging, tracing, and production metrics.

🧠 AI Engineer Intern — Simmspace.AI · Apr 2026 – Present · Remote

• Building Simmspace IDE and SimmSilos — a platform where admins assign branches/tasks and developers code in isolated silos via a custom VS Code extension.
• Designed a scalable FastAPI backend with JWT auth + refresh flow, RBAC, token-authenticated sync APIs, Redis rate limiting, audit logging, and PostgreSQL.
• Built a security-focused AI risk layer scanning submissions for prompt injection, PII leakage, and hallucination signals; gated PRs via GitHub Actions and Docker.

🔬 Undergraduate Research Assistant — CSULB · May 2025 – Apr 2026 · Long Beach, CA

• Built multithreaded Java/C++ programs to study parallel execution and optimized CPU-bound workloads across thread pools.
• Diagnosed race conditions with thread-safe synchronization and automated benchmark suites in Python/Bash to detect performance regressions.

🤖 Software Engineering Fellow — Headstarter AI · May 2025 – Jul 2025 · Remote

• Built dynamic front-end interfaces with React and TypeScript and a customer analytics tool (Python + React) that cut data processing time by 30%.
• Collaborated with industry mentors on AI-powered, open-source-style projects.

👨‍🏫 Instructional Student Assistant — CSULB · Jan 2025 – Present · Long Beach, CA

• Designed interactive Python workshops, increasing student engagement by ~15%, and supported 200+ students with targeted feedback.
• Revamped a math support course (50+ new participants) and trained new tutors on teaching strategies.

Data Structures & Algorithms · Object-Oriented Programming · Systems Programming · Digital Logic & Assembly · Discrete Mathematics · Intro to Software Engineering

• 💼 LinkedIn: linkedin.com/in/aryan-lakhani4443
• 🧑‍💻 GitHub: github.com/Aryan4443
• ✉️ Email: aryannlakhani21@gmail.com

If you're working on something in AI systems, safety/security, or full-stack — I'd love to connect or collaborate!