UCLA · Computer Science
Arjen Singh. Learning systems thatsee, reason & build.
I connect software to real engineering and research problems—turning point clouds, neural networks, and autonomous agents into tools that hold up outside the lab.
An afternoon at the Met. My working interests Novel AI Research LiDAR & Perception Neural Networks Agentic Software Engineering ✳ PROJECT 01 ↗
belief b(s) goal g π(b, g) Robotics · Planning under uncertainty
Combining offline dynamic programming with online Bayesian belief updates to study how a robot should navigate when it cannot confidently localize itself.
POMDPs Bayesian filtering Dynamic programming
1.44 → 0.29 Bottleneck MAP localization error
PROJECT 02 ↗
Supervisor PASS BLOCK conversation chatbot intervention LLM systems · Adversarial evaluation
A conversation-aware safety supervisor that evaluates dialogue history before a chatbot responds, tested across model sizes and moderation approaches.
LLM safety Multi-turn context Evaluation
99.5% 8B supervisor recall · with over-triggering
02 Experience
Engineering in practice. From sensing and tracking to enterprise AI and manufacturing software.
01
Sensing & autonomous systems
The Aerospace Corporation LiDAR Engineering Intern
Jun–Aug 2026
Work spanning long-range LiDAR characterization, telescope-array simulation, and resident space object (RSO) detection.
Selected work + Active/passive tracking and MHT versus PMDAF benchmarking. Automated beam profiling and LiDAR characterization. MCP servers and AI agents interacting with production systems. LiDAR Tracking Simulation AI agents
02
Software for manufacturing
Astra Space SWE Intern · Manufacturing Operations
Mar–Jun 2026
Engineering work on an ITAR-compliant LLM gateway serving 150+ enterprise users, alongside tooling for manufacturing operations.
Selected work + LiteLLM proxy and enterprise LLM agent workflows. Manufacturing bill of materials (mBOM)-backed tooling and real-time build/work-order visibility. Quality Control workflow improvements using Python, JavaScript, and GraphQL. Python JavaScript GraphQL LiteLLM
03
Enterprise AI workflows
Stealth Startup SWE Intern · Operations
Nov 2025–Jan 2026
RAG-based enterprise knowledge retrieval and operational task automation, in collaboration with Operations and Analytics.
Selected work + Business workflow automation and internal system integration. Enterprise AI workflows connecting knowledge retrieval with operational tasks. RAG Automation System integration
↳ Selected details omitted due to confidentiality.
03 A little about me
Curious about the model.Just as curious about the system. I’m a Computer Science student at UCLA, interested in what happens when machine learning meets the constraints of a real system.
My work spans robot navigation, LLM evaluation, and engineering tools. I’m especially drawn to problems where uncertainty is part of the design.
More about me