Research Overview

My research spans the intersection of programming language theory, systems engineering, and applied AI. I'm interested in how programs are structured, how memory is managed, how distributed systems are specified, and how AI can work reliably in resource-constrained environments.

Current Topics

  • Choreographic Programming with Perceus Memory Model
    Exploring ownership-based memory semantics in distributed, message-passing systems expressed as choreographies.
  • Metaprogramming in Choreographic Programming
    Compile-time protocol generation, verification, and transformation within choreographic languages.
  • Wyzer: LMVS Memory Model
    Linear-Mutable-Value-Semantics memory model for the Wyzer language, targeting GC-free, predictable systems-level performance.

Applied AI & Systems

  • Aethelix
    Probabilistic root-cause isolation in satellite telemetry streams.
  • Anveshak
    Offline artifact recognition and RAG for archaeology.
  • Agribot
    Edge AI vision models + quantized LLM for crop disease diagnosis.

Publications

Aethelix: A Physics-Based Causal Inference Framework for Real-Time Satellite Fault Detection

Atiksh Sharma • Zenodo (2026)

An open-source framework that encodes satellite subsystem failure mechanisms into a causal graph for rapid probabilistic fault isolation.

DOI Link