Aayush Joshi

Role: Associate Consultant / AI Platform Engineer

Organization: Alteryx

Project: Enterprise AI Automation & Evaluation Systems – Alteryx Designer Platform

Duration: 2025 – Present | [Remote]

Tech Stack: Python, Gemini Models, Vertex AI, DeepEval, LangGraph, LangSmith, PyTest, Playwright, Apache Airflow, Allure Reporting, XML Parsing, Slack Integrations

Responsibilities & Contributions:

• Autonomous Agent Engineering: Engineered an enterprise-grade AI automation framework utilizing Gemini-based agents to autonomously execute complex workflow-building challenges driven by multi-variable problem statements and dynamic scenario datasets.
• Multi-Agent Orchestration Pipelines: Built an end-to-end LLM orchestration pipeline integrating advanced design patterns, including LLM-as-a-User for scenario simulation and LLM-as-a-Judge for deterministic output validation.
• Workflow Reasoning & XML Diagnostics: Designed and implemented the Workflow Building Effectiveness (WBE) Judge, an advanced evaluation system performing structural XML-level analysis to map expected vs. actual transformations while detecting missing or incorrect implementation steps.
• Trajectory & Execution Analysis: Developed intelligent step-level evaluation logic and execution-tracing systems via LangGraph and LangSmith to validate agent decision-making behavior, conversational memory, tool selection, and non-deterministic workflow paths.
• LLMOps Observability & Reporting: Built advanced observability and enterprise reporting layers, customizing pipelines using DeepEval, Allure Dashboards, and Slack channel integrations to provide automated failure diagnostics, execution-debugging visualizations, and real-time status alerts.
• Cross-Functional Standardization: Collaborated closely with platform architects, data scientists, and core engineering teams to define formalized evaluation rubrics, scoring metrics, and AI workflow quality standards.

Key Achievements:

• Advanced Evaluation Automation: Delivered a proprietary WBE logic engine capable of performing automated root-cause analysis, sequence tracking, and normalized scoring for structurally equivalent but distinct transformation paths.
• Scalable Testing Infrastructure: Designed modular, scalable evaluation utilities for multi-scenario execution and workflow failure-cause extraction, future-proofing Alteryx's enterprise AI testing infrastructure.
• Production-Grade LLM Observability: Successfully combined agent trajectory analysis with enterprise reporting pipelines, supporting Alteryx's claims of delivering "secure, robust, and highly observable" agentic AI architectures and DAG-based automation systems.

Role: R&D Backend Engineer

Organization: Momentive Software (Community Brands)

Project: Gift Entry & Transaction Management – GiveSmart Donor CRM Tech Stack: Python, Flask, SQLAlchemy, Angular, PostgreSQL, CircleCI, Algolia, AWS

Responsibilities & Contributions:

• Enhanced the Donor CRM platform to support complex gift-types (In-Kind, Stock, Property) enabling nonprofits to track non-cash gifts just like cash donations.
• Designed and implemented dynamic form logic and validation flows to synchronize Fair Market Value, Gift Amount, and matching-gift eligibility.
• Led backend API work for the /transaction-mart module, ensuring transaction data flowed seamlessly into donor profiles, reporting pipelines and financial systems.
• Integrated wealth-based donor segmentation, custom-field templates and householding logic aligned with GiveSmart’s next-generation donor management features.
• Collaborated closely with product, QA and frontend teams to deliver UAT-ready builds, build feature flags, automate regression checks and ensure backward compatibility.

Key Achievements:

• Delivered enhancements that streamlined gift entry time by 85% for in-kind/stock donations.
• Implemented modular backend components reusable for other fundraising pipelines (events, peer-to-peer, auction) future-proofing the platform.
• Ensured compliance with data-security, audit-trail and donor recognition requirements supporting GiveSmart’s claims of “secure, scalable” donor-tooling.

Role: Python/Django Developer

Organization: MDP Infra Pvt. Ltd. (Client: MPMKVVCL – Madhya Pradesh Madhya Kshetra Vidyut Vitran Co. Ltd.)

Project: Quality Monitoring Portal -qcportal.mpcz.in

Tech Stack: Python, Django, Django REST Framework, PostgreSQL

Responsibilities & Contributions:

• Designed and implemented scalable backend systems using Python and Django, ensuring high performance and reliability.
• Developed RESTful APIs, integrating third-party services and ensuring seamless data flow.
• Conducted code reviews, identified performance bottlenecks, and optimized code for efficiency.
• Collaborated closely with cross-functional teams to meet unique and demanding requirements, delivering complex web applications that exceeded expectations.
• Worked as a Python Developer with client side of MPMKVVCL (Madhya Pradesh Madhya kshetra Vidyut Vitran Co. Ltd) i.e State Power Generation Company outsourced by MDP INFRA India Pvt. Ltd for the Commercial website development of MPMKVVCL for their Vendor/ NABL/ TKC registrations and LAB equipment onboarding.

Role: Python Developer & Automation Engineer

Organization: Tudip Technologies Pvt. Ltd., Pune

Project: TV2-TEST-AUTOMATION Framework Development

Tech Stack: Python, PyTest, OpenCV, Selenium, Computer Vision, Jenkins

Responsibilities & Contributions:

• Employed as a Python Developer and Automation Engineer, leading the development of the TV2-test-Automation-Framework. The framework specializes in testing the functionalities of Set-Top Boxes (STBs) and evaluating streamed content from platforms such as Netflix, Crave, and Amazon Prime.
• Leading the incorporation of advanced Computer Vision Technologies like Optical Character Recognition (OCR), Template Matching, and Motion Detection into the framework. Using Python packages to improve automation and streamline testing processes for enhanced efficiency.
• Contributing expertise to the augmentation of the testing framework's capabilities, particularly in the realm of STB functionalities and streamed content. Ensuring seamless compatibility with leading media platforms through continuous development and refinement.

Role: Associate Software Engineer

Organization: Webdunia.com (India) Pvt. Ltd., Indore

Project: OFFLINE-VOTING-CMS Development for Viacom-18 Tech Stack: PHP (Laravel), JavaScript, MySQL

Responsibilities & Contributions:

• Led the implementation of backend integrations for a prominent Indian OTT platform, guaranteeing a seamless and efficient operation of connections behind the scenes.
• Developed a user-friendly Content Management System (CMS) on the Laravel Framework for the VOOT OTT platform. This CMS played a crucial role in enabling offline voting, enhancing user engagement, and contributing to a more interactive experience.
• Played a key role in the creation and improvement of a CMS tailored for the Bulletin news app. This involvement focused on optimizing functionalities, usability, and overall user experience.
• Collaborated in enhancing the features and capabilities of the VOOT OTT platform through effective CMS integration, contributing to its position as a leading service provider in India.
• Applied technical expertise to streamline and improve backend processes, ultimately contributing to the overall efficiency and performance of the OTT platform's content management and delivery systems.