All My Reviews on Upwork

The images below are reviews from jobs I got from my Upwork account, which show that I have experience working as an AI engineer remotely.

This project, recognized by the client for "Excellent skills in AI development," involved the creation of a sophisticated web application called the "AI Truth Engine." Built with Streamlit, this tool leverages a suite of top-tier Large Language Models (LLMs) to perform complex analysis tasks, showcasing advanced AI engineering and research capabilities.

The platform features two primary modules:

1) Shadow Ban Detector: An efficient tool that processes lists of names from uploaded files (.csv, .xlsx) to identify potential online shadow banning. It queries multiple LLMs—including models from OpenAI, Anthropic, Google, and Groq—in parallel to generate a comprehensive downloadable report on the prominence and visibility of individuals.

2) Argument Analyzer: An interactive interface where users can input text and select from various analytical frameworks, such as "Logical Analysis Engine," "Evidence Scorer," and "Causal Reasoning Evaluator." The tool provides multi-faceted critiques and dissections of the argument from the perspective of different AI models.

Technologies Used: Python, Streamlit, Langchain, Pandas, OpenAI (GPT-4), Anthropic (Claude 3.5 Sonnet), Google (Gemini Pro), Groq (Llama 3.1, Mistral, Gemma2).

This project involved developing a targeted application for a Python programming study, for which the client praised my "Very good communication, delivered on time." The primary goal was to create a functional and reliable tool using the Flask web framework to facilitate the study's objectives.

The work focused on building a straightforward and effective backend that met all the specific requirements for the research participants. The successful and timely completion of this project underscores my ability to deliver quality Python-based solutions while maintaining clear and consistent communication with stakeholders.

Technologies Used: Python, Flask.

Praised by the client as "a pleasure to work with," who "always gave more than he was asked," this project involved the development of a sophisticated AI Brand Strategist. The client noted my proactive approach in the "fast-moving world of AI Agents," highlighting my ability to independently learn and implement new technologies to meet project goals.

The AI Brand Strategist is a powerful automated system built with CrewAI that orchestrates multiple specialized AI agents to perform a comprehensive brand analysis. The system is composed of several crews, each with a distinct role:

Brand Analysis Crew: Evaluates the brand's positioning, values, and engagement.

Competitor Analysis Crew: Analyzes the strengths, weaknesses, and strategies of competitors.

Competitor Reviews Crew: Mines customer reviews for sentiment and actionable insights.

Market Research Crew: Researches market trends, size, and the regulatory environment.

Audience Analysis Crew: Analyzes customer feedback, sentiment, and preferences.

Strategy Synthesis Crew: Consolidates all insights into a cohesive and actionable brand strategy.

The project leverages a suite of custom tools for in-depth analysis of brand perception, competitor intelligence, and customer insights. The final output is a series of detailed markdown reports, providing a comprehensive overview of the brand's strategic landscape.

Technologies Used: Python, CrewAI, OpenRouter (for LLM access), Exa API, Apify, Pydantic, python-dotenv.

Praised by the client as a "great professional" for delivering a "Knowledge Graph within a very short time frame," this project focused on building an advanced GraphRAG pipeline.

Using LangChain, I processed unstructured text to extract key entities and relationships, which were then structured and stored in a Neo4j Knowledge Graph. This graph-based approach enhances the Retrieval-Augmented Generation (RAG) process, allowing for more contextually aware and accurate information retrieval by leveraging the rich connections within the data.

Technologies Used: Python, LangChain, Neo4j.

This project, recognized by the client for "Good Work," involved creating a sophisticated, local AI troubleshooting assistant using LangChain and LangGraph.

The application processes user queries, including text and error screenshots, through an intelligent workflow. It features a robust Retrieval-Augmented Generation (RAG) pipeline that queries a local ChromaDB knowledge base using an Ollama-powered LLM. Key components include an OCR service for image processing and a critical security gate that detects and rejects any input containing Personally Identifiable Information (PII) to ensure data privacy. When a solution is found, it is presented to the user for approval; otherwise, the issue is automatically escalated.

Technologies Used: Python, LangChain, LangGraph, Ollama, ChromaDB, Hugging Face Embeddings, Streamlit.

Send me a DM on Contra