Ansh

Key Skills

Software

AWS SageMaker

Top Subject Matter

Voice AI (ASR/TTS)

LLM fine-tuning

edge NPU deployment

Top Data Types

Audio

Image

Document

Top Task Types

Fine-tuning

Function Calling

Segmentation

Classification

Data Collection

Freelancer Overview

Machine Learning Engineer III — HP Inc. Brings 7+ years of professional experience across complex professional workflows, research, and quality-focused execution. Core strengths include PyTorch, LangChain, and OpenAI GPT. Education includes Master of Science, Ohio State University (2024) and Bachelor of Technology, University of Mumbai (2022). AI-training focus includes data types such as Audio, Image, and Document and labeling workflows including Fine-tuning, Function Calling, and Segmentation.

Labeling Experience

Staff Machine Learning Engineer — SROT AI

Audio

Architected an end-to-end machine learning and business intelligence platform for predictive maintenance and anomaly detection across robot fleets. Built ML pipelines from real-time telemetry ingestion through feature engineering, model training, and model serving. Established MLOps best practices including experiment tracking and CI/CD to support iterative model deployment. • Ingested real-time telemetry streams via Kafka and engineered features • Developed serving endpoints for inference and operational BI dashboards • Built analytics visualization for robot telemetry KPIs and model outputs • Implemented MLflow/W&B tracking and CI/CD using GitHub Actions to reduce deployment cycle time

2026 - Present

Machine Learning Engineer III — HP Inc

AudioFine-tuning

Performed edge-native generative AI work by quantizing ASR and TTS models for deployment on edge NPUs. Implemented DoRA fine-tuning for voice-focused LLM adaptation to improve linguistic nuance capture. Led development of Voice AI platform components supporting speech recognition and wake word detection use cases. • Quantized ASR/TTS models and tuned for low-latency inference • Applied DoRA to fine-tune voice-oriented language behavior • Optimized multimodal voice-language models using hardware-aware acceleration • Developed Text-to-Speech, Speech-to-Text, and Wake Word detection modules

2025 - 2026

Graduate Research Associate — Ohio State University

ImageSegmentationClassification

Designed model architectures for fine-tuning image segmentation and classification systems in agricultural and ecology contexts. Worked on adapting vision models to domain-specific imagery for digital agriculture and animal ecology applications. Focused on training-ready pipeline setup and model behavior improvement for the target tasks. • Fine-tuned segmentation and classification models for domain imagery • Supported digital agriculture and animal ecology model adaptation • Prepared and validated training/inference architecture for the vision tasks • Iterated on model performance through experimentation

2024 - 2025

Student Research Assistant — Ohio State University

AWS SageMakerDocumentData Collection

Enabled AI tooling as part of an intelligent automation reference architecture for infrastructure management. Focused on preparing and integrating AI capabilities to support the automation workflow using the project's underlying data and models. Contributed to making AI components usable within the broader infrastructure management automation system. • Integrated AI tooling into the reference architecture • Supported intelligent automation for infrastructure management • Assisted with model/workflow readiness for operational use • Worked within a development team on AI-enabled automation

2024 - 2024

AI Engineering Intern — HP Inc

AudioFunction Calling

Built and orchestrated AI agents for system diagnostics and learning experiences using LLM APIs. Developed a multi-agent math tutor that combined voice input with computer vision to support problem-solving workflows. Used LangChain with hosted LLM APIs to automate multi-step analysis and tutoring behavior. • Created a Multi-Agent Orchestration System with LangChain and OpenAI/Claude APIs • Integrated voice and computer vision components into a math tutoring agent • Designed analyzer and actor-critic problem-solving roles for the tutoring system • Applied prompt and orchestration logic to drive automated diagnostics and learning

2024 - 2024

Education

Work History