About me

I’m a computer science student focused on deep learning research. I work on language and multimodal models, with a particular interest in how they learn, adapt, and behave.

See my research and experiments →

Current research

October 2026 — Present

San Jose State University

San Jose, CA

Researcher — Research under Prof. Mark Stamp

  • Studying whether Alice and Bob can learn a keyed communication scheme that resists a newly trained, ciphertext-only Eve.
  • Built controlled PyTorch experiments comparing CNN, MLP, BiLSTM, and Transformer attackers across model capacity and training budgets.
  • Early results show that fresh attackers can recover plaintext information above chance, with capacity and optimization especially affecting MLP leakage.

PyTorch · Neural Networks · Cryptography · Experimental Evaluation

Experience

Work and education

May 2026 — August 2026

A10 Networks

San Jose, CA

Deep Learning Engineer Intern - AI Firewall Team

  • Architected and trained TrooperViz for multimodal safety evaluation across text, image-text, multilingual, and multi-turn inputs, scaling PyTorch training across 8 GPUs and 160,157 examples.
  • Fine-tuned a vision-language model and merged it with a text-only guardrail using SLERP; achieved 91.83% single-turn and 90.24% multi-turn accuracy on heldout multimodal benchmarks across 21 safety policies, beating tested public guardrails by up to 16.4 percentage points. Works similarly better on Multilingual benchmarks too.
  • Built data and evaluation pipelines with Pandas and Polars, generating 85,000 synthetic conversations from a 340,473-row data pool and enabling interactive inference through Gradio and vLLM.

Python · PyTorch · SFT · SLERP · vLLM · Hugging Face · W&B · CI/CD

October 2025 — Present

Routes Technologies

Remote

AI Engineer Intern

  • Architected an LLM-powered NL-to-SQL pipeline with few-shot prompting, input normalization, 470+ synonym mappings, parameterized SQL sanitization, session memory, and Weights & Biases observability across 5 Azure ML endpoints.
  • Built an SVD collaborative-filtering recommendation system with scikit-learn and an LLM-based taste profiler, generating updated user preference profiles through automated nightly ETL jobs and achieving 86.7% Precision@5 and 89.8% Recall@10 on a 6-user, 87-recipe benchmark.
  • Engineered a multi-source recipe ingestion pipeline using Scrapy, social-media APIs, LLM classification, language filtering, and Faster-Whisper transcription to process video and web-based content across 5 source integrations.

Python · SQL · Azure ML · scikit-learn · Sentence Transformers · W&B · Scrapy

May 2025 — August 2025

Dreamable Inc.

San Francisco, CA

AI/ML Engineering Intern

  • Fine-tuned a Qwen 7B model using TensorFlow, Hugging Face, and Lambda, applying LoRA to reduce training cost and memory usage while delivering a domain-specific question-answering model.
  • Evaluated model configurations across multiple hyperparameters through systematic experimentation, tracking training and evaluation metrics with Weights & Biases to optimize model performance.
  • Developed an AI outreach agent using LangChain, Exa.ai, and LLM APIs to automate messaging workflows, reducing manual outreach time and improving response rates while documenting processes in Notion.

Python · PyTorch · TensorFlow · Hugging Face · LoRA · GCP · LangChain · W&B

Education

San Jose State University

B.S. Computer Science · GPA 3.88/4.00 · Expected May 2027

  • Dean’s List
  • Coursework: Object-Oriented Programming, Advanced DSA, Computer Architecture, Information Security, Machine Learning, Software Engineering

San Francisco State University

Computer Science (transferred) · 2023–2025

  • VP of AI Club & Tech Lead at SparkSF
  • Hosted SFHacks for more than 400 attendees
  • Dean’s List Honoree

Leadership & involvement

Leadership, teaching, and community building at SFSU and SJSU

Vice President AI Club @ SFSU

San Francisco State University

Hosted events, guest speakers, lectures, networking events for the club members

Tech Lead SparkSF

San Francisco State University

Built and managed the website for the club to help with organizing events

Logistics Team Member SFHacks

San Francisco State University

Helped organize and host a hackathon for 400+ attendees.

Teaching Assistant CS 215

San Francisco State University

Teaching assistant for Professor Duc Ta.

Advanced Data Structures & Algorithms

CodePath

Learning about overall idea as well as deep algorithm level of Advanced data structures and algorithms

Projects

Open-source repositories and model weights

LLM Firewall for Agentic Tool-Calling

Low-latency inline defense for prompt injection. A GPT-2 attacker loop reduced bypass from 23.1% to 3.72%; BERT with LoRA adds about 20 ms latency.

theHelper — AI Research Assistant

Production RAG with FAISS, LangChain chunking, cross-encoder reranking, local observability, and CI-gated QA.

Kanting — Video RAG System

Indexes YouTube transcripts with Whisper and returns precise clip timestamps through semantic search.

AdaptRoute

Task-aware routing from a query to specialized models.

GatorGPT

A 63M-parameter transformer for consumer GPUs, using Grouped Query Attention and RoPE.

How LLMs Are Made

Annotated implementations of GPT-2, DeepSeek MoE, and Kimi from scratch.

Hugging Face weights

MedAssistGPT 303M & 401M

Medical-domain transformers trained on PubMed, using SwiGLU, GQA, and RoPE.

Hugging Face

Qwen2.5-0.5B SFT+DPO 0.5B

A chat model fine-tuned with SFT and Direct Preference Optimization.

Hugging Face

Llama-3.2-3B OpenHermes 3B

QLoRA fine-tuning on filtered OpenHermes conversational datasets.

Hugging Face

StableLM 1.6B SFT+GRPO 1.6B

Aligned with GRPO on PKU safety preferences.

Hugging Face

Skills & credentials

Tools I use and things I’ve completed (B = Basics)

LLM engineering

Transformers · Pretraining · Fine-tuning · Inference · Alignment · Benchmarking · Multi-GPU training · SFT · DPO · GRPO · PPO · LoRA/PEFT · TRL · vLLM · Quantization · Vector databases

ML & NLP

PyTorch · TensorFlow · scikit-learn · Weights & Biases · Neural Networks · Reinforcement Learning · LangChain · FAISS · Whisper · BART · Sentence Transformers · Pandas · NumPy

Backend & cloud

FastAPI · Flask · Node.js · Docker · Azure ML · Microsoft Azure · GCP (B) · Lambda (B) · PostgreSQL (B) · MySQL · MongoDB · Scrapy · Ngrok · CI/CD (B)

Programming

Python · SQL · Java · JavaScript · C++ (B) · Bash · R · HTML/CSS · Git · Linux

Hackathons

  • SacHacks Med - U.C. Davis
  • Sparthill Hacks - SJSU
  • CalHacks 12.0 — Palace of Fine Arts, SF
  • MCP AWS Agentic Challenge — AWS Builder Loft, SF
  • SacHacks — Virtual
  • HackMerced — UC Merced
  • Cal Hacks 11.0 — San Francisco

Certificates

  • AI Memory: LLM Memory Systems — LinkedIn
  • Fine-Tuning for LLMs: Beginner to Advanced — LinkedIn
  • Model Context Protocol (MCP) — LinkedIn
  • Introduction to Generative AI — Google Cloud
  • Introduction to Web Development — UC Davis
  • Programming in Python — University of Michigan
  • Special Theory of Relativity — Stanford University
  • Calculus through Data & Modelling (×4) — Johns Hopkins

Hobbies

Outside of deep learning

Astrophotography

I photograph planets, star clusters, galaxies, and nebulae using long exposures and image stacking.

View my astrophotography collection →

A star named after me

There is also a registered star carrying my name.

View the star record →

Contact

For fine-tuning work, alignment evaluation, model safety, or just to say hello

The easiest way to reach me is by email.

kunjcr2@gmail.com