Production-style AI systems portfolio covering retrieval, multi-agent workflows, predictive scoring, anomaly detection, meeting intelligence, and an HR assistant application.

This repository is an index and review guide for the six runnable projects. It is not a seventh standalone application.

Open index.html for a portfolio-facing landing page designed for GitHub Pages. It summarizes the six runnable projects, links to each repo and demo guide, and gives 3-minute, 15-minute, and 30-minute reviewer paths.

Live site: https://adityansh-chand.github.io/ai-engineering-portfolio/

GitHub Pages setup:

Open this repository on GitHub.

Go to Settings > Pages.

Set source to "GitHub Actions".

Save the setting.

Push to main or run the Deploy GitHub Pages workflow manually.

Open the generated Pages URL after deployment finishes.

Manual repository setup is required once: GitHub Pages must be enabled in the repository settings and configured to use GitHub Actions as the source. The workflow does not require secrets or credentials.

Loading

Project Role Current runnable surface Verification enterprise-rag-knowledge-system Retrieval and grounded answer pipeline FastAPI /query, /metrics, SQLite event store, Compose/K8s config Locally tested, smoke-tested, Docker/K8s config statically validated, Docker image build validated in CI ai-proactive-customer-operations Multi-agent customer decision workflow FastAPI /decide, /metrics, SQLite event store, Compose/K8s config Locally tested, smoke-tested, Docker/K8s config statically validated, Docker image build validated in CI ai-incident-detection-platform Operational anomaly scoring FastAPI /score, /metrics, SQLite event store, Compose/K8s config Locally tested, smoke-tested, Docker/K8s config statically validated, Docker image build validated in CI ai-sales-intelligence-engine Account propensity scoring FastAPI /score, /metrics, SQLite event store, Compose/K8s config Locally tested, smoke-tested, Docker/K8s config statically validated, Docker image build validated in CI autonomous-meeting-intelligence Transcript structuring FastAPI /analyze, /metrics, SQLite event store, Compose/K8s config Locally tested, smoke-tested, Docker/K8s config statically validated, Docker image build validated in CI ADAAS Flutter HR assistant and Node HR backend Flutter app, secured backend, Mongo persistence, Compose/K8s config Locally tested, smoke-tested, Docker/K8s config statically validated, Docker image build validated in CI

GitHub Actions validates service image builds on push and pull requests without pushing images to a registry. Cluster deployment is still pending environment-backed cloud verification for all six runnable projects.

Python service pattern:

With the server running, use a second terminal:

Enterprise RAG uses a named eval runner:

ADAAS:

For demo paths and sample assets, see DEMO.md.

Project README API docs Env example Tests and eval Docker/Compose Kubernetes CI enterprise-rag-knowledge-system Yes Yes Yes Yes Yes Yes Yes ai-proactive-customer-operations Yes Yes Yes Yes Yes Yes Yes ai-incident-detection-platform Yes Yes Yes Yes Yes Yes Yes ai-sales-intelligence-engine Yes Yes Yes Yes Yes Yes Yes autonomous-meeting-intelligence Yes Yes Yes Yes Yes Yes Yes ADAAS Yes Yes Yes Yes Yes Yes Yes

Start at index.html or the GitHub Pages site for the 3-minute overview.

Use DEMO.md to pick one runnable service and follow its exact smoke path.

Open each target repo README for purpose, quickstart, API surface, deployment status, and remaining gaps.

Inspect docs/ARCHITECTURE.md, docs/API_FLOWS.md, and docs/TRADEOFFS.md for system-level reasoning.

Treat this repository as the portfolio index only; the six linked repos are the runnable projects.

Expect local demos, tests/evals, static deployment config validation, and CI Docker image builds; cloud deployment and production data remain pending.

Open DEMO.md and choose one service from the demo matrix.

From the Workspace folder or a clone of the target repo, start enterprise-rag-knowledge-system:

In a second terminal, run:

Inspect examples/requests/query.json and examples/responses/query.json.

Repeat the same pattern for any scoring, orchestration, transcript, or HR assistant repo of interest.

Project Status Labels enterprise-rag-knowledge-system locally tested, smoke-tested, Docker config statically validated, Docker image build validated in CI, cloud deployment pending, needs production data ai-proactive-customer-operations locally tested, smoke-tested, Docker config statically validated, Docker image build validated in CI, cloud deployment pending, needs production data ai-incident-detection-platform locally tested, smoke-tested, Docker config statically validated, Docker image build validated in CI, cloud deployment pending, needs production data ai-sales-intelligence-engine locally tested, smoke-tested, Docker config statically validated, Docker image build validated in CI, cloud deployment pending, needs production data autonomous-meeting-intelligence locally tested, smoke-tested, Docker config statically validated, Docker image build validated in CI, cloud deployment pending, needs production data ADAAS locally tested, smoke-tested, Docker config statically validated, Docker image build validated in CI, cloud deployment pending, needs production data

Core retrieval reasoning backbone using semantic chunking, reranking, and confidence scoring. https://github.com/Adityansh-Chand/enterprise-rag-knowledge-system.git

Explicit multi-agent DAG orchestration implementing planner to specialist to action workflow. https://github.com/Adityansh-Chand/ai-proactive-customer-operations.git

Production HR assistant integrating RAG reasoning with real-time API data. https://github.com/Adityansh-Chand/ADAAS.git

Predictive ML pipeline for customer intelligence scoring. https://github.com/Adityansh-Chand/ai-sales-intelligence-engine.git

Anomaly detection system for operational intelligence. https://github.com/Adityansh-Chand/ai-incident-detection-platform.git

Structured transcript understanding pipeline for summaries, decisions, and action items. https://github.com/Adityansh-Chand/autonomous-meeting-intelligence.git

Typed request/response boundaries for APIs.

Domain-specific sample data instead of generic demo CSVs.

Focused tests that assert real system behavior.

Lightweight evaluation scripts with explicit accuracy or structure metrics.

Docker entrypoints that run FastAPI services through uvicorn.

Deterministic local fallbacks where external providers are optional.

Optional X-API-Key auth on non-health data endpoints.

Request IDs, safe error responses, and JSON metrics endpoints.

SQLite event persistence for Python services; MongoDB persistence for ADAAS.

GitHub Actions CI across tests, evals, and container builds.

Capture and link final screenshots or short recordings per system.

Add a shared API contract document for common request tracing.

Add common logging and metric naming conventions.

Add managed cloud deployment targets and release environments.

Adityansh Chand

AI Software Engineer specializing in multi-agent systems, retrieval engineering, LLM architecture, and machine learning pipelines.