Vaibhav Savala

Software Engineer

Washington, DC, United States6 yrs 9 mos experience

Key Highlights

Expert in Java and Python with extensive backend development experience.
Proven track record in Machine Learning and Deep Learning projects.
Strong skills in AWS and data engineering for healthcare applications.

Stackforce AI infers this person is a Backend-heavy Fullstack Developer with a focus on Healthcare and SaaS industries.

Contact

Skills

Core Skills

PythonAwsData EngineeringGcpJavaSoftware DevelopmentData Analysis

Other Skills

Fine TuningDeep LearningRetrieval-Augmented Generation (RAG)HealthcareAmazon Web Services (AWS)Amazon Elastic MapReduce (EMR)Apache SparkTextractMavenKubernetesSnowflakeAmazon Relational Database Service (RDS)OpensearchAWS TextractMicrosoft Azure

About

As a passionate engineer, I bring a blend of expertise in Java and Python, honed through years of building robust applications from the ground up. My proficiency extends to working with both Spring and Flask frameworks, allowing me to create scalable and efficient systems tailored to specific needs. My experience isn’t just limited to application development; I am equally skilled in managing databases, whether SQL or NoSQL, ensuring data integrity and optimal performance. I have also worked closely with customers to deploy, debug, and fix issues, provide customer support, and present new features, ensuring a seamless user experience and continuous improvement of the products. What excites me is the dynamic field of Machine Learning and Deep Learning. I have channeled this enthusiasm into developing cutting-edge projects, including deepfake detection and exploring the intricacies of adversarial ML. These projects not only highlight my technical skills but also reflect my commitment to staying at the forefront of technological advancements.

Experience

6 yrs 9 mos

Total Experience

1 yr 8 mos

Average Tenure

1 yr 7 mos

Current Experience

Hilabs

SDE II, AI/ML

Nov 2024 – Present · 1 yr 7 mos · Washington DC-Baltimore Area · On-site

Part of contractsAI team
Development and scaling of data science services for the ContractsAI product, leveraging Docker, Amazon EKS and SQS, resulting in a 60% improvement in information retrieval speed.
Optimized OCR processing for health contracts using batching techniques with Amazon Textract, improving searchable PDF conversion and reducing retrieval latency by 65%.
Automated the contract information extraction process, reducing turnaround time by 99% and significantly enhancing efficiency.
Developed a Retrieval-Augmented Generation (RAG) pipeline using OpenSearch to store and retrieve document chunks, achieving 95%+ accuracy in complex data retrieval.
Fine-tuning of Document Understanding Transformer (Donut) model for creating bounding boxes for accurate text retrieval using label studio for creating coco dataset.
Migrated from an native deployment of MistralAI 7B to Amazon Bedrock Llama-3 90B, increasing attribute extraction accuracy from 61% to 95% while improving scalability and performance.
Roster Automation
Implementation of metadata-driven rule engine for multi-market provider data processing, enabling scalable and configurable rule execution across 20+ U.S. health-plan markets.
Built reusable contract crosswalk rule templates that reduced rule creation time by 40–60% and ensured consistent logic validation across waves.
Engineered dynamic SQL-based validation rules with JSON-driven condition objects, improving accuracy and reducing manual remediation cycles.
Automated ingestion of crosswalks and lookup tables using Python/SQLAlchemy, reducing onboarding time for new markets by 30%+.

PythonFine TuningDeep LearningRetrieval-Augmented Generation (RAG)HealthcareAmazon Web Services (AWS)+15

Strative

Software Developer

Sep 2024 – Nov 2024 · 2 mos · San Francisco Bay Area · Remote

Developed the backend for a Retrieval-Augmented Generation (RAG) Enablement Platform, utilizing Elasticsearch and Pinecone as vector databases to enhance data retrieval.
Designed and implemented custom Llama 3.1 models integrated with Elasticsearch, enabling context-aware language models for healthcare and legal data applications.
Achieved an 80% reduction in data vectorization time through distributed and batch processing techniques.
Applied advanced quantization techniques to improve retrieval speed and overall system efficiency.
Successfully implemented Chain of Thought (CoT) reasoning models using OSS models such as Llama 3.1, Med_Llama, and GPT-3.5, achieving accuracy comparable to the MedRAG paper's benchmarks.
Leveraged extensive datasets for retrieval, including Textbooks, Statpearls, PubMed, and Wikipedia, using advanced models like Contriever, MedCPT, BM25, and Specter.
Developing a modified Contextual RAG, MedPrompt techniques for beating the accuracy on benchmarking datasets using OSS LLMs.
Optimized processing in a distributed environment utilizing L4 and T4 GPUs, significantly enhancing retrieval speed and computational efficiency.

PythonFastAPIGoogle Cloud Platform (GCP)Elastic Stack (ELK)Distributed SystemsRetrieval-Augmented Generation (RAG)+6

Digital.ai

Software Developer R&D

Jul 2020 – Aug 2022 · 2 yrs 1 mo · Remote

Part of execution team of the Digital.ai’s SaaS cloud in an agile environment.
Implemented SSL/TLS protocols to enhance security for on-premise cloud setups, and establishing regional proxy server infrastructure to efficiently manage multi-region customer access and ensure seamless connectivity.
Authored Ansible scripts for efficient deployment and upgrades of cloud-based builds.
Building a pipeline to integrate data from the product to elasticsearch cluster and creating kibana dashboards.
On-premise automation orchestration, troubleshooting and deployment of enterprise cloud in multi-region capacity and guiding developers to migrate UI test projects from a third party framework to Appium open-source framework.
Led the project for direct bug reporting to VersionOne, improving issue tracking and resolution speed from automation suites.
Performed debugging, and testing of the latest Appium open-source server/client to seamlessly integrate it into the product.
Successfully migrated an Appium automation test suites from Java 8 to Java 17, ensuring compatibility, optimizing performance, and also integrating Appium Open Source Software automation suites to our cloud testing product, achieving 100% test coverage
on supported commands.

JUnitJavaSoftware DevelopmentAnsibleDockerSpring Boot+6

Experitest (now digital.ai)

Technical Engineer

Apr 2019 – Jul 2020 · 1 yr 3 mos · Remote

Software DevelopmentJenkinsiOSAndroidCloud ApplicationsPostgreSQL+2

Power2sme

Software Developer

Jun 2017 – Apr 2019 · 1 yr 10 mos · Gurgaon, Haryana, India

I have worked as a Java developer using Spring framework. I also have experience in developing JavaScript and AngularJS applications.
Power2SME Website:
Developed a search engine on Power2SME News data using ElasticSearch.
Developed a query based search algorithm for searching various products of the company using ElasticSearch
Used Gsuite APIs (Directory API) to communicate between server to server applications.
Content Management System(CMS)
A role based mechanism for uploading most recent Tenders of raw materials to the power2SME website.
Create-> Save->Upload . Full stack development of the feature.
JSON Web Token Service(JWT)
Developed a jar which acts as a layer for authorising the access of backend APIs using Spring Security and Spring Boot. It is a token based Authentication.
Employee Portal
Developed a role based full stack web application for internal company users where the users can access the different applications/products from here. This has been developed in Java, spring boot, CSS, JavaScript, Thymeleaf technologies.
Logging Jar
Developed a common jar for logging each request and its response and store the logging object in NoSQL database. Technologies used were ActiveMq, MongoDb.
NBFC Portal:
Worked on NBFC portal, a role based mechanism for loan approval process. My major contribution was in migrating the web application which was in SpringMVC before to a springboot application and day to day product enhancements.

Unit TestingJavaMavenSoftware DevelopmentPL/SQLJenkins

Qa skills consultants private ltd

Big Data Analyst

May 2016 – Jul 2016 · 2 mos · Kolkata

I have worked as a Big data analyst intern here. I have worked mainly on R and mySQL.
During my internship, I have worked on the following projects:
Customer Life Time Value and Churn Prediction: Developed a model with 80% accuracy on the test data.