Ayush Sharma

Machine Learning Engineer

Atlanta, Georgia, United States10 yrs 6 mos experience

Most Likely To SwitchAI Enabled

Key Highlights

10 years of experience in AI and Data Science
Reduced telecom customer service costs by 12%
Boosted SLA of services by 35%

Stackforce AI infers this person is a Data Science and Machine Learning expert with a focus on AI-driven solutions across multiple industries.

Contact

mail2sharmaay@gmail.com LinkedIn

Skills

Core Skills

Machine LearningData Science

Other Skills

APIsAmazon Web Services (AWS)Analytical SkillsAnomaly DetectionApache AirflowApache SparkAzure DatabricksAzure cloud servicesBig DataBusiness Intelligence (BI)D3.jsData AnalysisData EngineeringData MiningData Transformation

About

Results-driven and certified Azure, DeepLearning.AI, Machine Learning, Data Science professional with 10 years of experience in designing and developing advanced analytical and AI driven solutions across multiple industries – Telecom, Manufacturing, Supply Chain, Education, Relocation, Geoscience, and IT consulting. Proficient in managing cross-functional teams to produce innovative solutions. Deep understanding of machine learning, data analytics, and NLP, and technical expertise in building scalable solutions, data engineering pipelines, gathering insights from data. Notable achievements include decreasing telecom customer service costs by 12%, avoiding 30% of manufacturing drive train failures, and boosting SLA of services by 35%.

Experience

10 yrs 6 mos

Total Experience

2 yrs 1 mo

Average Tenure

6 yrs 5 mos

Current Experience

Ivoyant

Machine Learning Engineer

Jan 2020 – Present · 6 yrs 5 mos · Atlanta, Georgia, United States

Generative AI Engineer Lead at a Client Project
Led the design and development of an AI-driven orchestration workflow generator, using OpenAI, Llama, Claude LLMs and Azure cloud services to design data pipelines from user prompts. Managed the team of data scientists, testers, developers, and SREs.
Developed APIs for prompt engineering, inferencing, and performance benchmarking using ROUGE and BLEU metrics.
Implemented a robust RAG pipeline for structured and unstructured data, leveraging vector database and embeddings.
Fine-tuned LLM using QLoRA PEFT technique and confined accurate outputs using dynamic Pydantic schema steering.
Senior Machine Learning Engineer at a Client Project
Planned and implemented Machine Learning workflows for Telecom industry to decrease costs and enhance customer satisfaction. Devised KPIs and illumined insights on performance of customer service. Deployed distributed data infrastructure on Kubernetes cluster to process real-time events and warn abnormalities driven by anomaly detection system employing Apache Beam.
Built Customer Churn Prediction model which helped retaining 12% of Telecom customers for any given month.
Enterprise Data Architect at a Client Project
Designed and developed a real-time stream processing pipeline using Kafka streams to replace the inefficient legacy data transformation pipelines. Analyzed business data to identify trends and patterns to generate insights into capacity planning and resource allocation to help the business make informed decisions.
Senior Data Engineer at a Client Project
Managed design and development of predictive maintenance model for foreseeing discrepancies and wearing of components in drive train. Restructured product dimensional data models manufacturing plant to foster relational connections. Devised and led data transformations for both batch and streaming and maintained a fault-tolerant and highly available data infrastructure.

TensorFlowMicrosoft AzureData MiningApache SparkAnalytical SkillsKubernetes+30

Edge geoscience inc

Data Scientist

Jun 2019 – Dec 2019 · 6 mos · Houston, Texas Area

Conceptualized and developed drilling hazard predictor and lithological facies classification model for the largest oil drilling well site in West Permian Basin, Texas. Led the R&D wing of AI team to devise statistical workflows for geoscience operations.
Leveraged TensorFlow & DeepLearning AI algorithms to reduce drilling fluid consumption and cost by 25%
Developed real-time monitoring dashboard using D3.js intended to optimize drilling and raise potential hazards.
Streamlined real-time WITSML well-log data using Apache Airflow DAGs for analytical and visualization solutions.

Microsoft AzurePythonData MiningApache SparkAnalytical SkillsElasticSearch+25

Deloitte india

Risk Advisory Consultant

May 2017 – May 2018 · 1 yr · Mumbai Area, India

Predicted vendor defaults using Neural Networks & forecasted commodity prices using time series modeling.
Analyzed Procure to Pay process of India’s biggest FMCG to identify process gaps, control failures and frauds.
Automated a control monitoring platform analyzing 75+ Procurement Analytic tests extended on Tableau.
Illuminated insights on Revenue Leakages and Procurement frauds concealed in $300M+ annual spends.
Developed an interactive dashboard, showcasing 75+ risk controls associations using Association Rules Mining.
Modeled an image recognition Deep Learning model to classify vehicle types at toll plazas.

SAPSASTensorFlowPythonData MiningApache Spark+12

Zaya learning labs

Sr. Machine Learning Engineer

Sep 2015 – May 2017 · 1 yr 8 mos · Mumbai Area, India

Built an education platform driven by recommendation engine - enhancing personalized skill set of 5k+ kids.
Analyzed app data to calculate user engagement metrics to highlight acquisitions, retention and churns outs.
Developed Teacher Assistant NLP model - generating written content using semantics & induced grammar.
Envisioned Learning Maps by mapping raw documents to meaningful topics using latent semantics and NLP.
Developed an English learning game facilitating behavioral analytics and hierarchical knowledge maps, featuring speech and handwritten character recognition on mobile.
Clustered the app users to gain pedagogical insights based on their performance learning pattern.

PythonData MiningAnalytical SkillsElasticSearchKubernetesDeep Learning+14

Grownout solutions

Data Scientist

Jun 2015 – Aug 2015 · 2 mos · New Delhi Area, India

Built a statistical model to predict job switching pattern of working professionals using 40M+ LinkedIn profile.
Highlighted professional insights like trending skills, orgs, switching pattern of cities and similar profile cluster.

PythonAnalytical SkillsLinear RegressionData ScienceSQL

Cybage software

Software Engineer

Jul 2014 – Jun 2015 · 11 mos · Pune Area, India

AthenaHealth - Developed back end patient and hospital report generation.
HotelsCombined - To extract and maintain (ETL) the data of all the competitive hotels for parameters such as Price, Ratings, Reviews, Address, etc.

PythonSQL