Ashesh Kaji

Expected to defend an honors thesis on: "Effect of metal exposure from vapes on brain imaging"

• Consulting on AI tooling pipelines for the same workplace, spanning LLM deployment, data transformation, and production MLOps

• Designed pipelines for custom data transforms and fine-tuning strategies for varied data lifecycle requirements
• Led development of major products: LLMs, SLMs, statistical mapping models, and semantic database memory retrieval

• Designed and implemented pipelines for data extraction and chat interface creation using Large Language Models (LLMs) and vector databases
• Developed a Retrieval-Augmented Generation (RAG) product that went into production, involving integration with Azure and serverless deployment of locally-hosted LLMs
• Conducted architecture design, code implementation, rapid versioning, and production testing

• Studied relationship of peripheral iron levels and NAFLD on neurodegenerative disorders using UK BioBank data
• Explored effect of metal exposure from vapes on MRI imaging of subjects from the ABCD study
• Applied Python, CLI, NumPy, Pandas, and Statsmodels API with epidemiology methodology

Deep reinforcement learning for autonomous driving systems using CARLA simulator with real-world sensor feeds. Studied policy gradients, DQN variants, and sim-to-real transfer.
github.com/ashesh8500/fp185 ↗

Research on portfolio allocation as layered system optimization: walk-forward validation, regime detection, and attractiveness scoring across screened pilot universes with provenance-ledger tracking.
Interactive research demo ↗  ·  github.com/ashesh8500/fractal ↗

Fast, local-first pipeline for large photo/video libraries: content-deduplication with SHA-256, legacy container transcoding to MP4 with hardware encode, and batched cloud upload via rclone/rsync. Built in Rust with a terminal UI.
github.com/ashesh8500/mediasync ↗

Built an app to transform Apple Health export data into analyzable CSVs with visualization for studying relationships between blood glucose, exercise, and other biomarkers.
github.com/ashesh8500/apple_health_export ↗