Arham Doshi
CS + Linguistics @ UT Austin · AI/ML + SWE
Published NLP researcher — ACL & CoNLL 2025.
Publications
Research
What I bring
Skills & highlights
Research
- ACL 2026 — LLM faithfulness vs. safety in medical RAG
- CoNLL-2025 — SES prediction via Reddit + NLP
- RARS — LSTM-based Hindi-English Jain scripture translation
- UT Austin Dell Medical — Speech Disorders & Technology Lab
Leadership
- Convergent (AI/ML Club) — Software Engineer; won $200 in competitive project showcase
- Nonprofit founder — Jain Enlightened (500+ users)
- Top-3 pitch at PM Accelerator program
- VEX Robotics World Championship qualifier
Career
Experience
2026
Research Assistant
UT Austin Dell Medical · Wang Lab (Speech Disorders & Technology Lab)
Research Assistant
UTSW · Wenqi Lab
Built a recurrent LLM-memory framework for MIMIC-IV diagnosis prediction, achieving 0.80 Acc@1 and eliminating LLM-judge reward hacking.
2025
Undergraduate Research Assistant
UT Austin — Discourse Lab (Junyi Jessy Li)
Faithfulness vs. Safety (ACL 2026) — studied how frontier LLMs handle counterfactual medical evidence, revealing critical safety gaps in RAG systems. Built evaluation pipelines using Azure OpenAI across multiple experimental conditions.
Research Assistant
Baylor College of Medicine — Hayden Lab
Transcribed and analyzed 4,000+ verbal responses from epilepsy patients. Extracted acoustic features in Praat; improved transcription model by 12%.
2024
Research Assistant, Co-Author
Southern Methodist University — CS Dept.
Co-authored CoNLL-2025 paper on SES prediction via Reddit + NLP. Preprocessed 30k+ entries with NLTK, spaCy, LDA, and transformer models.
AI Software Engineering Intern
PM Accelerator
Designed college match platform with ML recommendations and speech-to-text. Delivered top-3 final pitch.
Research Assistant – Meme Analysis (NLP/ML)
University of Texas at Dallas
Conducted a para-review analyzing meme dissemination using NLP and machine learning.
Selected work
Tiny fraction of my work
Stack
Technical Skills
- Languages
- Python
- C/C++
- Java
- TypeScript
- JavaScript
- Go
- Rust
- SQL
- R
- Bash
- MATLAB
- Deep Learning
- PyTorch
- TensorFlow
- JAX
- Transformers
- CNNs
- LSTMs
- CUDA/Triton
- DeepSpeed
- FSDP
- Accelerate
- LLMs & NLP
- Hugging Face
- LoRA/QLoRA
- RLHF/GRPO
- vLLM
- RAG
- FAISS
- LangChain
- LLM agents
- prompt engineering
- spaCy
- NLTK
- BERT/RoBERTa
- Sentence-BERT
- Machine Learning
- scikit-learn
- scikit-survival
- XGBoost
- LightGBM
- feature engineering
- PCA
- K-means
- cross-validation
- Data
- pandas
- NumPy
- SciPy
- Polars
- Dask
- Spark
- Matplotlib
- Seaborn
- Plotly
- Airflow
- dbt
- ETL/ELT
- Cloud & MLOps
- AWS (SageMaker, Bedrock, Lambda, S3)
- GCP
- Azure
- Docker
- Kubernetes
- Terraform
- W&B
- MLflow
- Backend
- FastAPI
- Flask
- Node.js
- REST APIs
- GraphQL
- WebSockets
- PostgreSQL
- MongoDB
- Redis
- Tools & HPC
- TACC (Lonestar6, Vista)
- SLURM
- Apptainer
- Ray
- Linux
- Git
- Jupyter
- Conda
- LaTeX
Background
Education
University of Texas at Austin
B.S. Computer Science + Linguistics
Focus: AI/ML, NLP, LLM Evaluation · Austin, Texas
Get in touch
Let's build something.
Open to research collaborations, internships, and AI/ML opportunities.
© 2026 Arham Doshi