Lalu Prasad Lenka
Verified Expert in Engineering
Data Scientist and Software Developer
Dublin, Ireland
Toptal member since July 12, 2022
Lalu is a senior data scientist at AWS and has a master's degree in data science from Trinity College Dublin. He has 5+ years of experience formulating research-driven approaches to solve challenging business problems using data and state-of-the-art machine learning algorithms. Lalu has deployed scalable ML solutions on cloud platforms like AWS and Azure and on-premise Kubernetes clusters. He presented compelling insights to many stakeholders, helping them make data-driven decisions.
Portfolio
Experience
- Data Science - 6 years
- Python 3 - 5 years
- Machine Learning - 5 years
- Deep Learning - 4 years
- Data Analysis - 4 years
- Natural Language Processing (NLP) - 3 years
- PySpark - 3 years
- Generative Pre-trained Transformers (GPT) - 3 years
Preferred Environment
Linux, Visual Studio Code (VS Code), Databricks, Jupyter Notebook, MacOS, Slack, Agile, Agile Sprints, Jira, Amazon Web Services (AWS)
The most amazing...
...project I've developed is an ML solution for a fast-fashion client that recommended favorable styles, forecasted demand, and increased profitability by 15%.
Work Experience
Senior Data Scientist
Amazon Web Services (AWS)
- Helped AWS build scalable and self-service go-to-market strategies.
- Leveraged generative AI to build personalized and curated guidance for AWS customers.
- Built personalization for a go-to-market platform from the ground up. Tackled business challenges like lack of training data, ambiguous business problems, and stakeholder alignment.
- Created a customer segmentation model to send tailored messaging to customers for a client in the advertising and marketing industry.
- Developed an optical character recognition (OCR) solution for an insurance industry client using Amazon Textract. The deployed solution extracted information from PDF documents to automate the client's claims processes.
- Built a recommender system for a retail client using Graph Neural Networks trained on top of data stored in the graph database AWS Neptune. The solution helped improve the precision metric by 30%.
- Engineered a search engine on top of Amazon OpenSearch, which supports full-text and semantic search. It leveraged OpenSearch's vector database abilities to support GenAI Retrieval Augmented Generation (RAG) use cases.
Data Scientist
AAYS Analytics
- Extracted, aggregated, and analyzed large data sets to provide actionable insights; also created intuitive visualizations to convey those results to a broader audience.
- Analyzed profit erosion for a finance client and discovered adverse cost components which helped optimize existing revenue streams.
- Developed and deployed an intelligent supply chain solution for a fast-fashion client that helped the client maintain optimal stock levels for favorable clothing styles and increased earnings.
- Contributed to building the data infrastructure for client organizations on Azure, including setting up a data lake, ETL (data engineering) pipelines, and machine learning pipelines.
- Acted as a data scientist to build and operationalize reliable and scalable machine learning pipelines for data preparation, model training, and prediction at scale. Deployed data pipelines on the Azure cloud platform.
- Led client meetings and presented compelling findings and a story for the "why" of these findings to a wide range of stakeholders with insightful visualizations using Power BI reports.
Data Scientist
Aptus Data Labs
- Served as a data scientist to partner with clients to understand their business pain points and design analytical solutions to address those; also helped clients use their organization's data to drive strategic business decisions.
- Focused on data preprocessing, machine learning modeling, and the operationalization of ML models.
- Developed and deployed an LSTM-based (named entity recognition) model for a pharma client that helped reduce manual efforts by 90%.
- Developed and deployed an inventory optimization platform that used hybrid time series models for long-term forecasting and demand sensing. This helped the client maintain optimal inventory for products and plan demand fulfillment.
- Developed and deployed a deep learning pipeline for a manufacturing client that performs text localization and recognition, helping reduce human error and operations costs by 40%.
Data Scientist
Aptus Data Labs
- Worked as a data science intern on time series analysis and text analytics projects.
- Implemented, for a Fortune Global 500 oil-and-gas company, a proof of concept for a supply chain optimization project by creating a time series model to forecast the load(oil, gas) requirement at different ports based on historical data.
- Developed, for a multinational pharmaceutical company, text-analyzing software to migrate thousands of documents into a different format. It helped them reduce the operational cost of merger by 5%.
- Created tools for a sanity-check-like document comparison tool to visually analyze the difference in two almost similar documents. Successfully automated the whole process and reduced manual effort to a staggering 1-2% of the initial effort.
Machine Learning Expert
Tata Consultancy Services
- Worked on a project called "Image Attribute Extraction" which includes extraction of text from product images and populating specific attributes with extracted text.
- Developed a Keras model for text recognition using connectionist temporal classification loss.
- Developed a CNN-RNN based neural network to detect text in product images that helped the team build a more robust text extraction.
Experience
Profit Erosion Analysis
Afterward, I discovered loss-making consumer pack types and adverse cost components. The next task I did was to perform a prescriptive analysis, and then I suggested business actions that optimized existing revenue streams by 18%.
Soon after, I performed prescriptive analytics and used time series forecasting to project the sales, major cost components and predicted the potential losses if no action taken. Finally, I created a "What if Tool" to prescribe the next best business action.
Fast Fashion Intelligent Supply Chain Solution
The solution helped the client maintain optimal stock levels for all SKUs and increased profitability by 15%.
Chemical Named Entity Recognition
The model was developed on the IUPAC dataset and achieved an F1 score of 0.85.
Machine Screen Text Recognition
Text extraction on an image was done in two independent steps—detection (region proposal network) and recognition using CNNs. I achieved a mAP (mean average precision) of 0.56.
Demand Forecasting and Demand Sensing
This helped the client maintain optimal inventory for all products and plan demand fulfillment.
IoT Streaming Analytics Platform
Autonomous Car Parking
https://github.com/Lplenka/Autonomous-Car-Parking• Reinforcement learning – soft actor-critic (SAC)
• Imitation learning – behavior cloning
• Neuroevolution – genetic algorithm
The above algorithms were used to build three different AI agents that tried to control the steering and acceleration to park the vehicle correctly. The performance of these agents was compared for multiple simulations.
Car Damage Detection Using Semantic Segmentation
https://github.com/Lplenka/Car-Damage-DetectionGiven a pic of the damaged car, find which part is damaged. The parts can be either the rear bumper, front bumper, headlamp, door, or hood.
Prescriptive Guidance Using Generative AI and RAG
https://pathfinder.aws.com/Recommender System for Retail Client
Customer Support Chatbot using AI Agents
Education
Master's Degree in Data Science
Trinity College Dublin - Dublin, Ireland
Bachelor's Degree in Computer Science
Odisha University of Technology and Research - Bhubaneswar, India
Certifications
Specialization in Statistics
Coursera
Deep Learning Specialization
Coursera
Skills
Libraries/APIs
Scikit-learn, TensorFlow, PySpark, Spark ML, LSTM, PyTorch, SpaCy, Natural Language Toolkit (NLTK), Keras, TensorFlow Deep Learning Library (TFLearn)
Tools
Slack, AWS CodeDeploy, Kafka Streams, Microsoft Power BI, Amazon Elastic Container Service (ECS), Jira, Microsoft Excel, Git, Amazon SageMaker, AWS IAM, GitHub, Spark SQL, ChatGPT, MongoDB Atlas
Languages
Python 3, Python, SQL, R
Paradigms
Agile, Azure DevOps
Platforms
Jupyter Notebook, Linux, Azure, Docker, Kubernetes, Databricks, AWS Lambda, Apache Kafka, Amazon Web Services (AWS), Visual Studio Code (VS Code), MacOS, Windows, AWS Elastic Beanstalk
Storage
Graph Databases, MySQL, Azure SQL Databases, Amazon S3 (AWS S3), Neo4j, Azure Cosmos DB
Frameworks
LangGraph
Other
Machine Learning, Data Science, Artificial Intelligence (AI), Time Series, Large Language Models (LLMs), Amazon Machine Learning, Semantic Search, Embeddings from Language Models (ELMo), Retrieval-augmented Generation (RAG), Data Engineering, Machine Learning Operations (MLOps), Chatbots, Prompt Engineering, English, Sentiment Analysis, Data Analytics, Data Visualization, Inventory Management, Deep Reinforcement Learning, Deep Learning, Data Analysis, Computer Vision, Natural Language Processing (NLP), Statistical Methods, CI/CD Pipelines, Demand Planning, Serverless, MLflow, Azure Data Factory (ADF), Azure Data Lake, Genetic Algorithms, Statistics, Agile Sprints, Optical Character Recognition (OCR), Text Mining, Predictive Modeling, Predictive Analytics, Data Queries, Web Scraping, DeepAR, Agile Data Science, Text Recognition, Text Detection, Time Series Analysis, Generative Pre-trained Transformers (GPT), LangChain, Amazon Neptune, eCommerce, OpenAI, Supply Chain Management (SCM), Model Deployment, Hugging Face, Kalman Filtering, Forecasting, Prescriptive Analytics, Prescriptive Modeling, Stakeholder Management, Personalization, Recommendation Systems, Generative Artificial Intelligence (GenAI), Knowledge Bases, AI Chatbots, Amazon Bedrock AgentCore, Graph Neural Networks (GNNs), Scalable Vector Databases, AI Agents, FastAPI, Amazon Route 53
How to Work with Toptal
Toptal matches you directly with global industry experts from our network in hours—not weeks or months.
Share your needs
Choose your talent
Start your risk-free talent trial
Top talent is in high demand.
Start hiring