Looking for a job

Abhay Kumar

@abhaykumar13

Message

AI Engineer at Zenwork Inc. who built a tax-focused AI Copilot and cut LLM inference costs by 40%.

India

Message

At Zenwork Inc., I built the backend and AI system for a multi-domain AI Copilot supporting US taxation, product workflows, and issue resolution. Its hybrid retrieval and conversation-state management reduced LLM inference costs by 40%.

I also built an AI calling system that handled 100K+ calls with a 98% success rate, combining real-time speech processing, LLM-based reasoning, and backend workflows. I developed an automated bot-training pipeline that reduced manual training effort by 45%.

Earlier, I prepared pharmaceutical datasets for ML workflows at Customerinsights.ai and developed HMM-based pulse prediction models at Apexplus Technologies. My projects include FinSentry AI, a data-processing and anomaly-detection system, and graph-based analysis to detect circular trading.

Experience

Work history, roles, and key accomplishments

Current

AI Engineer

Current

Zenwork Inc.

Dec 2025 - Present (10 months)

Built backend and AI systems for a multi-domain AI Copilot, including hybrid retrieval, AI calling, and automated training pipelines. Developed an LLM Router and OCR-based payment processing, reducing inference cost by 40% and manual training effort by 45%.

Python fastAPI LangGraph BM25 Embeddings MiniLM LLM RAG Speech Processing OCR Server Sent Events RBAC LLM routing AI Copilots Backend Development

Data Science Intern

Customerinsights.ai

Aug 2025 - Oct 2025 (2 months)

Prepared and cleaned large-scale pharmaceutical datasets for ML workflows, performing data validation, anomaly detection, and preprocessing. Used Python, SQL, and Excel for exploratory data analysis and transformation.

Python SQL Excel Data Cleaning Data Validation Anomaly Detection Pre processing Exploratory Data Analysis

Internship Manager

Office of Career Services, IIT Hyderabad

Apr 2024 - May 2025 (1 year 1 month)

Led 80+ student coordinators and managed internships for 500+ students across 200+ companies. Organized Japan Career Day with 18 Japanese firms and drove a 40% YoY increase in internships.

Leadership Coordination Event Management Student Management

Python and Machine Learning Intern

Apexplus Technologies

Jan 2024 - Jun 2024 (5 months)

Developed HMM-based pulse prediction models and used DBSCAN to cluster radar signal patterns. Built preprocessing pipelines and automated testing workflows for signal attenuation evaluation.

Python Machine Learning HMM DBSCAN Signal Processing Pre processing Testing Automation

Application Developer Intern

Suzuki Motor Innovation Center

Oct 2022 - Mar 2023 (5 months)

Developed an IoT-based water quality monitoring system with secure AWS microservices for real-time analysis. Created AWS data pipelines and real-time dashboards with alerts, contributing to early MVP delivery.

IoT AWS Microservices Data Pipelines Real time Dashboards Alerts