01. The Fellowship
AI Research Fellowship
The AI Research Fellowship is Redex’s most selective program designed for Masters and PhD students who want to work on applied AI problems that matter, not academic exercises that never leave the lab.
Over 12 weeks, you’ll embed with our R&D team and work on real research challenges: computer vision systems that process enterprise data, predictive analytics models that inform business decisions, and natural language processing pipelines that power our AI agent platform.
This isn’t a typical internship. You’ll have the freedom to explore, the mentorship to guide you, and the infrastructure to run experiments at scale. We provide GPU access, production datasets (anonymized), and direct collaboration with senior AI engineers who’ve shipped models to production. If your research produces publishable results, we’ll support your paper submission.
02. Research Areas
Where you'll contribute
Choose a primary research area based on your expertise. Cross-pollination between areas is encouraged.
01
Computer Vision & Document AI
- Document understanding and intelligent extraction
- Visual question answering for enterprise documents
- OCR enhancement and layout analysis
- Multi-modal reasoning across text and images
Tools & Frameworks
- YOLO
- LayoutLM
- OpenCV
- Hugging Face
- PyTorch
02
Predictive Analytics & Time Series
- Demand forecasting for enterprise clients
- Anomaly detection in operational data
- Customer behavior prediction models
- Time series foundation models evaluation
Tools & Frameworks
- scikit-learn
- XGBoost
- Prophet
- TensorFlow
- Pandas
03
LLM & Agentic AI Systems
- RAG pipeline optimization and evaluation
- Multi-agent orchestration frameworks
- LLM fine-tuning for domain-specific tasks
- Hallucination detection and mitigation
Tools & Frameworks
- LangChain
- LlamaIndex
- OpenAI API
- Anthropic API
- Weights & Biases
03. The Journey
From hypothesis to publication
No ping-pong tables. No unlimited PTO that nobody takes. A typical week involves client workshops, solution design sessions, proposal writing, and delivery check-ins.
Week 1-2
Research Onboarding
Literature review, problem scoping, and hypothesis formulation. Meet your research mentor, access GPU infrastructure, and define your 12-week research plan.
Week 3–5
Experimentation Phase I
Build baseline models, run initial experiments, and iterate on your approach. Weekly check-ins with your mentor. Present early findings to the R&D team.
Week 6-8
Experimentation Phase II
Refine your approach based on Phase I results. Scale experiments. Start writing up methodology and preliminary results. Collaborate with engineers on production integration.
Week 9-10
Results & Analysis
Finalize experiments, conduct ablation studies, and analyze results. Begin drafting your research paper or technical report.
Week 11-12
Paper & Presentation
Complete your research paper. Present findings to Redex leadership and client stakeholders. Discuss publication strategy and potential full-time research roles.
04. What You’ll Gain
Beyond the lab
The skills and resources that make this fellowship different from a university research assistantship.
Applied Research Skills
Bridge the gap between academic research and production AI. Learn to build models that work in the real world, not just on benchmarks.
GPU Infrastructure
Access to production-grade MLOps pipelines, and enterprise datasets.
Industry Mentorship
Work directly with senior AI engineers who’ve shipped models to production at scale. Learn what academia doesn’t teach.
Production Exposure
See how your research translates to real products. Understand the engineering constraints that shape AI system design.
Research Network
Join a community of AI researchers and practitioners. Access to internal tech talks, paper reading groups, and industry connections.
05. Must-haves
Who should apply
This is a research-intensive program. We need strong foundations.
- Currently enrolled in a Masters or PhD program in Computer Science, AI/ML, Statistics, or related field
- Strong foundation in machine learning. You should be comfortable with gradient descent, loss functions, and model evaluation
- Proficiency in Python and at least one deep learning framework (PyTorch preferred, TensorFlow accepted)
- Experience reading and implementing research papers. You should be able to reproduce results from a paper
- Strong mathematical foundations (linear algebra, probability, statistics, optimization)
- Available for the full 12-week program
06. nice-to-haves
Bonus points
These will strengthen your application significantly.
- Published research papers (even workshop papers or preprints count)
- Experience with LLMs, RAG systems, or multi-agent frameworks
- Kaggle competitions or ML challenge experience
- Contributions to open-source ML projects
- Experience with MLOps tools (MLflow, Weights & Biases, DVC)
- Background in NLP, computer vision, or time series analysis