Lewati ke konten utama
Avinash Giradkar

Avinash Giradkar

Principal Application Engineer | Applied AI & Agentic Document Intelligence | Java, Python, OCI/AWS, Kubernetes

Talenta
New York, NY, United StatesAnggota sejak 9 Oktober 2026
Generative AILLMsRetrieval-Augmented Generation (RAG)LangChainOCRevaluationregression testingCodexJavaPythonSpring BootHelidonREST APIsKafkaSQLDB2RedisApigeeSwaggerOCIDockerKubernetesHelmPivotal Cloud FoundryJenkinsBambooGitMavenGradleCI/CD
·

Sewa orang ini

Kirim tawaran pekerjaan langsung ke kandidat ini

Generative AILLMsRetrieval-Augmented Generation (RAG)LangChainOCR

Tentang

Hands-on Principal Applications Engineer with 18+ years of software engineering experience delivering enterprise applications, cloud-native microservices, and reusable platforms. Combines enterprise Java expertise with Python-based AI development, including Generative AI, LLMs, OCR, retrieval-augmented generation (RAG), and agentic workflows. Technical experience spans Kafka, Kubernetes, Oracle Cloud Infrastructure (OCI), and AWS.

Leads architecture, implementation, and production delivery across distributed teams. Delivered document automation that reduced manual effort by 70% for order documents and 95% for bank-statement processing and led a telecom expense migration that eliminated $1.2M in annual licensing costs. Builds reliable systems through automated recovery, monitoring, and controlled deployment practices.

Pengalaman

Principal Applications Engineer at Oracle (2021-06 – Present)

  • Document+: Developed a reusable extraction, detection, and validation platform supporting bank statements, telecom invoices, and utilities, with extensibility to order documents (OD), purchase orders, contracts, and other document types.
  • Automated document evaluation with validation success/failure outcomes and auto-approval for documents that passed validation, reducing manual effort for OD documents by 70%.
  • Architected and led bank-statement processing using Helidon MP and Kafka, combining primarily deterministic extraction with OCR. Supported 50+ international banks and led AMER, EMEA, and APAC rollout, reducing manual entry and verification effort by 95%.
  • Implemented hybrid deterministic, LLM, and OCR processing for telecom invoices and utilities; delivered the telecom-invoice pilot using Python, Docker, Kubernetes, and OCI Generative AI.
  • Led the associated data-center telecom expense migration to Oracle Fusion-based systems, consolidating 5,500+ inventory records, supporting control over $250M in annual spend, and eliminating $1.2M in annual third-party licensing costs.
  • Prompt Builder: Conceived and architected a separate application for onboarding document types onto Document+; led implementation of an Oracle JET Template Configurator, UI API, and Prompt Builder services.
  • Extended bank-statement deterministic/LLM routing into template-driven Java and agent workflows. Enabled model selection, prompt lifecycle controls, role-based access, versioning, regression testing, and all-or-nothing publishing; used Codex to accelerate implementation.
  • Document Hub: Led architecture and early development of the intended Document+ successor and common intake point for Oracle documents. Designed classification, intent identification, and profile-driven extraction and validation, including signature detection requirements for OD documents.
  • AI Approval Advisor: Architected policy-based recommendations for requisition, procurement, and other requests using LangChain RAG. Designed human review to accept or reject recommendations, with automated approval of eligible requests planned as a future capability.
  • AgentVerse: Participated in the first team onboarded to another team's no-code agentic workflow platform; planned application migration and joint enhancements informed by onboarding findings.
  • Platform Reliability, Recovery, and Observability: Built shared Grafana monitoring and alerting across the application portfolio to track operational health and SLA risks; supported production under a 15-minute SLA.
  • Created Log Watcher, an agentic application monitoring Kubernetes logs and triggering pod restarts, persistent-volume cleanup, and custom remediation scripts to improve workload continuity.
  • Implemented Document+ recovery and retry mechanisms for stalled processing, supporting manual replay and application-directed corrective actions to automatically recover transactions.
  • Instrumented AI Approval Advisor to capture user actions on recommendations, enabling business teams to assess recommendation acceptance and policy-document effectiveness.
  • Technologies: Python, Java, JS, Helidon MP, REST APIs, Kafka, Docker, Kubernetes, OCI, OCI Generative AI, OCR, LLMs, LangChain, RAG, Grafana, Codex, Jenkins, Cucumber, Git, Oracle JET, SQL, Cline

Digital Solution Architect at SIDGS (2018-09 – 2021-05)

  • Architected enterprise API strategy and delivery roadmaps using Apigee, including requirements definition, solution validation, proof-of-concept delivery, and integration specifications.
  • Designed and implemented API security, OAuth2/JWT authentication, monitoring, logging, throttling, caching, rate limiting, monetization, and developer-portal capabilities.
  • Built Docker images and CI/CD pipelines with Jenkins, Maven, Groovy, Kubernetes, and Helm; developed Node.js extensions for Apigee Microgateway.
  • Built Swagger-based custom client SDK generation, Node.js Microgateway plugins, and Jenkins/Groovy/Maven CI/CD pipelines; deployed Docker services to Kubernetes using Helm.
  • Technologies: Apigee, Apigee Microgateway, Java, Spring Boot, Spring Security, REST, OAuth 2.0, JWT, Node.js, Swagger, Redis, Docker, Kubernetes, Helm, Jenkins, Groovy, Maven, Gradle, Git, Cucumber, Mocha, Chai, Sinon, nyc, WSO2, Drupal

Senior Systems Analyst at UST Global (2014-07 – 2018-09)

  • Modernized Vanguard applications into Java microservices on Pivotal Cloud Foundry; designed a transaction-processing framework, configurable DB2 SQL, Apache POI/SAX parsing, and reusable rules. Added Cucumber BDD tests and Sonar/JaCoCo quality checks.
  • Design and develop a framework which allowed us to process multiples of transaction, with each transaction having its own independent requirements. This design allowed us to deliver this critical business requirement with multiple stakeholders within the deadline.
  • Architected and Implemented custom parsing and file scanning logic in Apache POI and SAX parser. Implemented multiple complex business rules to be applied on a single transaction using abstracted rules implemented using reg-ex.
  • Technologies: Java, Spring Boot, REST/JAX-RS, DB2, SQL, Apache POI, SAX, AngularJS, JavaScript, Pivotal Cloud Foundry, Cucumber, Sonar, JaCoCo, Git/Bitbucket, Bamboo, Jenkins, Maven, Tomcat, IBM FileNet

Software Engineer at Infosys (2011-01 – 2014-08)

  • Provided onshore/offshore technical guidance and client coordination for Time Warner Cable. Integrated ICOMS/CSG billing systems and built Java SOAP services, JPA/Hibernate persistence, and JSP/Servlet interfaces.
  • Technologies: Java, EJB, Spring, JPA/Hibernate, HQL, SOAP/JAX-WS, JSP, Servlets, JavaScript, XML, WebLogic, WebSphere, JUnit, Selenium, SoapUI, YUI, DWR, JiBX

Software Engineer at Amicon Technology (2009-12 – 2010-12)

  • Built e-commerce and ERP solutions with Apache OFBiz and Opentaps, integrating Funambol, ProcessMaker, and Endicia to support client workflows.
  • Technologies: Java, Spring, Apache OFBiz, Opentaps, Funambol, ProcessMaker, Endicia, JUnit

Software Engineer at Tmaxsoft (2008-01 – 2009-01)

  • Migrated COBOL applications to C/C++ for OpenFrame on Unix and developed shell scripts and ANT build support.
  • Technologies: C/C++, COBOL, Tmax/OpenFrame, Unix, shell scripting, ANT

Pendidikan

M.S. in Computer Engineering – San Jose State University (2007)

B.E. in Computer Science and Engineering – Walchand College of Engineering (2004)

Keterampilan

Principal Application EngineerApplied AI & Agentic Document IntelligenceJavaPythonOCI/AWSKubernetes

Ulasan