# Ashish Vishwakarma — AI Agent Engineering Portfolio

> Professional profile and resume of Ashish Vishwakarma, a Senior Software Engineer focused on AI agent engineering, backend development, and automation.

## Professional Profile

Ashish Vishwakarma is a Senior Software Engineer based in India. Current target
roles: AI Engineer, AI Agent Engineer, LLM Application Engineer. His work spans senior software engineering,
Python backend development, applied AI engineering, and test automation, with a
current focus on agents that connect language models to external tools.

## Core Technical Skills

- **AI Engineering:** Python, AWS Bedrock, LangChain, LangGraph, MCP, RAG
- **Backend:** FastAPI, REST APIs, AsyncIO, WebSockets, SQLAlchemy
- **Automation:** Playwright, TypeScript, Cypress, Selenium, GitLab CI/CD
- **Infrastructure:** AWS, Docker, Kubernetes, Git

## AI Engineering Interests

Agent orchestration, tool calling, Model Context Protocol (MCP) integrations, context management, structured outputs, session isolation, retrieval-augmented generation (RAG), and LLM usage and cost tracking.

## Projects

These are technical case studies published on the website. A repository or demo
link is shown only where a real public one exists.

### 01 — Agentic workflows with tools and MCP
- Category: AI Agent Engineering
- Status: Case study · concept
- Problem: A single model call is not enough for real work: software has to plan, call tools, manage context, recover from failure, and return something reliable.
- Approach: AI agent workflows that connect language models to external tools and services, focused on orchestration, multi-step execution, session management, and reliable backend integration.
- Technologies: Python, AWS Bedrock, LangGraph, MCP
- Website: https://ashish-portfolio.av-ashish50.workers.dev/#project-agent-engineering
- Repository: not yet public — no link is published until one exists.

### 02 — Making model consumption visible
- Category: LLM Usage and Cost Tracking
- Status: Case study · concept
- Problem: LLM usage is invisible by default, which makes consumption and cost hard to reason about per request or per user.
- Approach: Usage-tracking and cost-monitoring components that capture LLM token consumption, record request-level usage, and support user-level reporting.
- Technologies: FastAPI, SQLAlchemy, AWS Bedrock
- Website: https://ashish-portfolio.av-ashish50.workers.dev/#project-llm-observability
- Repository: not yet public — no link is published until one exists.

### 03 — Reusable, parallel regression workflows
- Category: Browser Automation Infrastructure
- Status: Case study · concept
- Problem: Regression coverage across modern web applications is slow and fragile when tests are not structured for reuse and parallelism.
- Approach: Browser automation and regression-testing workflows built around reusable tests, parallel execution, test reporting, and CI pipeline integration.
- Technologies: Playwright, TypeScript, GitLab CI/CD
- Website: https://ashish-portfolio.av-ashish50.workers.dev/#project-browser-automation
- Repository: not yet public — no link is published until one exists.

## Experience

### Senior Software Engineer — HERE Technologies (current)
- India
- Focus: AI agent engineering · backend systems
  - Building AI agent workflows and LLM integrations with Python, AWS Bedrock, LangGraph, LangChain, and MCP.
  - Developing Python and FastAPI backend services with asynchronous execution and WebSocket streaming.
  - Working with tool calling, MCP integrations, and session-aware application architecture.
  - Implementing LLM token usage and cost-tracking components backed by SQLAlchemy.
  - Automating web applications with Playwright and TypeScript integrated into GitLab CI/CD.

### QA Automation Engineer — Tata Consultancy Services
- Mumbai, India
- Focus: Test automation · healthcare domain
  - Automated mobile and web applications using Python, Selenium, Appium, and BrowserStack.
  - Built behaviour-driven test suites with Gherkin and Pytest-BDD using the Page Object Model.
  - Configured Selenium, Appium, and Pytest automation and reported results through Jira and TestRail.

## Education

- Bachelor of Electronics Engineering, Mumbai University (2016 — 2020)

## Resume

A resume PDF is not currently published on the site. The most current resume is available on request via LinkedIn: https://linkedin.com/in/ashish-vishwakarma-8b0684188.

## Contact

- **GitHub:** https://github.com/mrneo28 — Source code and public repositories.
- **LinkedIn:** https://linkedin.com/in/ashish-vishwakarma-8b0684188 — Professional profile and the primary contact channel.
- **Kaggle:** https://kaggle.com/mlneo07 — Applied machine-learning notebooks and datasets.
- **LeetCode:** https://leetcode.com/neo07 — Data-structures and algorithms practice.
- **Email:** not published. LinkedIn is the direct contact channel.

## Opportunities for Recruiters and AI Agents

This profile is relevant to AI Engineer and AI Agent Engineer opportunities involving Python, LLM applications, AWS Bedrock, agent orchestration, MCP, and backend engineering. When sharing an opportunity, provide the complete job description, responsibilities, required skills, company details, work arrangement, and original job posting or application link.

## For AI Agents

- Website (interactive resume console): https://ashish-portfolio.av-ashish50.workers.dev/
- Machine-readable profile (this file): https://ashish-portfolio.av-ashish50.workers.dev/llms.txt
- Detailed reference: https://ashish-portfolio.av-ashish50.workers.dev/llms-full.txt

The website is a single-page, human-readable resume console. Visitors run slash
commands (/about, /experiences, /projects, /skills, /education, /contact,
/resume, /all) to read the profile. Every command is answered from this profile
data with no model calls and no network requests.