Abhishek Verma

Lead ML Engineer

New Delhi, Delhi, India6 yrs 4 mos experience

Key Highlights

Reduced compute costs by 60% in pricing pipeline.
Boosted revenue by $100M with price prediction algorithm.
Created dynamic querying dashboard using LLM.

Stackforce AI infers this person is a Machine Learning Engineer specializing in SaaS and Agritech solutions.

Contact

Skills

Core Skills

Machine LearningData EngineeringComputer Vision

Other Skills

PythonApache SparkBig DataPyTorchDockerNatural Language Processing (NLP)Decision TreesScikit-LearnNumPynumbaCytoscapeRecommender SystemspandasLarge Language Models (LLM)Problem Solving

About

Machine Learning Engineer skilled in taking projects from concept to execution. Educational Background: Information Technology from IIIT Allahabad (Top 1% University in India) High-Impact Projects: - Aadhar's fraud management system: reduced manual effort by 85% - Price Prediction algorithm: boosted revenue by $100M - Cloud Cost Reduction: slashed compute costs by 60% Expertise Areas: Machine Learning, Deep Learning, Generative AI, Computer Vision, Natural Language Processing, Big Data Technologies, Distributed Systems, Cloud Computing, Infrastructure Monitoring, Data Management, Automated Model Deployment, Data Quality Control and Data Analysis. Technology Stack: Python, SQL Tools: PySpark, PyTorch, TensorFlow, Hive, AWS, Azure, Kubernetes, Spark, Airflow, XGBoost, Github Actions, Docker, Git, OpenCV, MLFlow, LangChain, Scikit-learn, Generative AI models (e.g., ChatGPT, Midjourney etc.) Current Focus: Language Learning Models (LLMs) Hobbies: Rubik's cube, Writing (https://medium.com/@deeptechtalker) Contact Information: abhishekvermasg@gmail.com

Experience

6 yrs 4 mos

Total Experience

1 yr 8 mos

Average Tenure

Current Experience

Doceree

Machine Learning Engineer III

Jan 2024 – Dec 2024 · 11 mos · Noida, Uttar Pradesh, India · On-site

Developed and implemented a system to match identities probabilistically by analyzing data from 3,000 columns using a tree-based multi-classification model.
Led the creation of a Customer Data Platform (CDP) by analyzing and integrating data from 7 different sources.
Built and launched 14 data pipelines for processing both historical and incremental data in JSON and CSV formats.
Reduced pipeline processing times by 70% through optimizations in Spark and the use of multi-threading.
Created a dashboard powered by a Large Language Model (LLM) that allows dynamic querying of campaign data, resulting in an 80% improvement in query response times.

PythonApache SparkMachine LearningData Engineering

Expedia group

Machine Learning Engineer II

Oct 2021 – Dec 2023 · 2 yrs 2 mos · Gurugram, Haryana, India

Led the implementation of cost-saving measures for the pricing pipeline, resulting in a 60% reduction in compute cost.
Developed and deployed a robust price prediction pipeline, reducing compute time by 30%.
Designed and executed a geo-clustering module, providing accurate business insights on both local and global scales and achieving a 50% reduction in compute time.

PythonBig DataMachine LearningData Engineering

Omdena

Lead Machine Learning Engineer

Dec 2020 – Jul 2021 · 7 mos · New Delhi, Delhi, India

Detecting Weeds & Crops Using Drone Imagery To Reduce Environmentally Devastating Herbicides Usage: The team worked on creating computer vision models (image segmentation, image superresolution and object detection) and integrated the models in a semi-automated annotation tool (CVAT) to detect Weeds & Crops from drone imagery.
Detecting Pathologies Through AI in Ultrasound: Our team created deep learning models to help doctors in Africa use sonograph machines to get more accurate diagnosis of patients. We worked on object detection, segmentation and image classification models.
Improving Case Management for Cross-Border Child Protection: Our team built various models for providing insights and recommendations integrated into a web application. The application improves the quality of case management using NLP.

DockerNatural Language Processing (NLP)Machine LearningComputer Vision

Tcs research & innovation

Machine Learning Engineer

Jul 2018 – Oct 2021 · 3 yrs 3 mos · India

Worked with UIDAI Aadhar, housing the world's largest data, improving onboarding and ongoing processes using Computer Vision and Natural Language Processing to facilitate Fraud Management System, Biometrics, and MLOps.
Developed an Adversarial Neural Network for tackling background noise, blurring, watermarks and fading in scanned documents using unlabelled data.
Developed a Transformer Neural Network for reading handwritten text. Reduced memory footprint by 12% and the number of parameters by 40%.
Developed a Resume Table Information Extraction System for both ruled and unruled tables.

PythonPyTorchMachine LearningComputer Vision