Aditya Gidh

Advisory Software Engineer at IBM Research with 10+ years building scalable web platforms, developer tools, and cloud-native services. I drive initiatives spanning AI-powered developer experiences, production observability, and DevSecOps automation partnering across research, product, and engineering organizations to bring innovative solutions into production at scale. Previously at IBM Watson for Genomics and IBM Marketplace.

10+

years of industry experience

99.9%

uptime across 10+ microservices (Prometheus/Grafana)

8

reusable AI agent plugins authored

3,000+

users on IBM Carbon Design System rollout

Current Focus

AI & LLM Engineering

RAG pipelines, LangChain/LangGraph, vector databases, prompt engineering, AI-powered CLI tools. Contributor to beeai-framework, instructlab, granite-agent-cookbook. Jupyter notebooks featured at PyData London 2026 (570+ clones, 780+ unique users).

Platform & Observability

Prometheus/Grafana monitoring for 10+ microservices. DevSecOps with Tekton Secure Pipelines. Kubernetes/OpenShift deployments at scale.

AI-Ready Codebases

Pre-computed architectural intelligence for 61 services: materialized views of actual code designed for AI consumption. Inspired by Karpathy's wiki concept, extended with cross-service blast-radius analysis, known risks, and an MCP query interface. Now the context source an agentic delivery pipeline reads before writing a spec.

Agentic Delivery Pipelines

Developer Software Factory, in use across production pipelines: chaining DAM, GitHub, a spec-driven development CLI, and SPS into one AI-assisted path from a plain-language problem description to a sandbox-tested pull request, with a human validation gate preserved at every step. ROI measurement is the next milestone, not yet claimed.

Recent Publications & Talks

all publications →

Open Source

agent-readiness

author

reusable AI agent plugin for Claude Code, Codex, Gemini CLI

beeai-framework

contributor

AI agent framework

instructlab

contributor

LLM fine-tuning toolkit

granite-agent-cookbook

contributor

Agent patterns for IBM Granite

web-application-reference

contributor

Full-stack web app reference architecture

Community

AAIF Security & Privacy Working Grouptechnical contributor

Producing open threat models, shared taxonomy, and design patterns for trustworthy agentic AI - Agentic AI Foundation (AAIF)

Building Agents with Granite Workshoptechnical contributor

Contributed to the inception of this hands-on learning program delivered by the Granite team across IBM business units and technical organizations, and showcased at open-source, AI, and developer conferences worldwide.

Stack

languages  Python, TypeScript, JavaScript, Java, R

ai/ml      LangChain, LangGraph, RAG, FAISS, ChromaDB, Ollama, IBM Granite

frontend   React, Next.js, Astro, Angular, Carbon Design System

backend    Node.js, Express, PostgreSQL, MongoDB, Elasticsearch

infra      Kubernetes, OpenShift, Docker, Helm, Tekton, Prometheus, Grafana

full breakdown →

Elsewhere

linkedingithubemail