Generative AI Engineer · Serial Founder

Sai Prasad Muppala

AI Engineer Founding Product Engineer AI-Native SaaS Builder

I ship production AI end-to-end — agentic systems, RAG pipelines, and full-stack products — across public-sector, enterprise, and startup environments. 4+ years of execution, 10+ AI ventures, and a habit of turning hard problems into shipped product.

South Florida, USA Authorized to work in the U.S. M.S. Data Science · 3.89 GPA
Get in touch GitHub LinkedIn
Sai Prasad Muppala
10+AI
Ventures
4+Years
Shipping
0+
Years building production AI
0+
AI-native ventures founded & built
0
M.S. Data Science GPA · FAU
0
Gov & enterprise orgs shipped for
0×
NIH AI Hackathon — Winner
0
Patent · GCP-certified engineer
01 — ABOUT

Builder first, across every layer.

A serial founder and Generative AI engineer who moves fluidly from frontend to backend to model orchestration — translating ambiguous problems into polished, production-grade software.

I hold a Master's in Data Science & Analytics from Florida Atlantic University (3.89 GPA) and have spent 4+ years shipping AI where it has to actually work — inside the Florida Department of Health, the U.S. Army Corps of Engineers (via Smart Structures), the South Florida Water Management District, and Deloitte (Gilead Sciences, State of Colorado).

In parallel, I build. I've founded or co-founded 10+ AI-native ventures spanning media, careers, education, and developer tooling — from ClonA AI and Visra (AI dubbing and lip-sync for Indian cinema across 12+ languages) to Knuckler AI (AI career intelligence), GradeUS (AI grading), FinalDraft AI, and Sakura (AI-native workflow automation with billing built in). With SaSa Launch I co-led an AI SaaS built on cloud agents — multi-agent orchestration across the stack, from architecture decisions to rapid MVP execution.

My stack centers on agentic AI (LangGraph, CrewAI, MCP), RAG, and full-stack product (Next.js, FastAPI, Supabase), backed by real cloud, data-engineering, and ML foundations. My ties to Indian cinema and languages shape the dubbing and language-AI work. Y Combinator applicant, NIH AI Hackathon winner, and a patent holder.

Agentic AI Systems

Multi-agent pipelines, RAG, evals & observability for trustworthy LLM workflows.

Full-Stack Product

Polished surfaces to backend APIs — 0→1 SaaS with billing, auth & dashboards.

Data & Cloud

PySpark, BigQuery, Snowflake & event-driven pipelines on AWS / GCP.

Founder Mindset

Customer-driven shipping — feedback into measurable product, fast.

02 — EXPERIENCE

Where I've shipped.

Production AI and data engineering across public-sector, defense-adjacent, and enterprise programs — built to be secure, observable, and scalable.

State of Florida
Palm Beach County Health Dept
AI / ML Engineer
Dec 2025 — May 2026
Florida, USA
  • Designed AI-powered backend and analytics solutions to evaluate large-scale public-health submissions, automate reporting workflows, and deliver structured insights for leadership decision-making.
  • Built scalable data processing, cleansing, validation, and integration workflows using Python, SQL, Oracle, and cloud data sources, improving data accuracy, audit readiness, and operational reporting reliability.
  • Developed AI-assisted document-intake pipelines using OCR, NLP, structured parsing, and validation logic to extract key fields, classify submissions, identify missing data, and sharply reduce manual review effort.
  • Built mobile-first data-collection applications with Microsoft Power Apps, enabling field teams to capture reports, upload documents, and trigger automated backend workflows.
  • Designed interactive Power BI dashboards on Oracle and cloud sources delivering real-time KPIs, anomaly detection, operational trends, compliance metrics, and executive-level insights.
  • Integrated Microsoft Copilot, Python automation, Oracle pipelines, and the Power Platform to streamline reporting, documentation, approvals, and operational decision-making.
  • Architected integration-ready backend workflows using REST API patterns, structured data models, access control, audit logging, and event-driven automation for secure, scalable public-sector operations.
PythonOracleOCR / NLPPower AppsPower AutomateCopilotPower BIREST APIs
Knuckler AI
Stealth · Click AI Solutions
Founding AI Engineer
Aug 2025 — Dec 2025
Remote, USA
  • Designed and built AI-powered backend services and RESTful APIs using Python, FastAPI, Next.js, PostgreSQL, and a metadata-driven architecture for document intelligence, evaluation workflows, semantic scoring, and customer-facing AI features.
  • Shipped production GenAI and agentic capabilities with LangChain, LangGraph, CrewAI, MCP Server, GPT models, Claude, Gemini, DeepSeek, RAG pipelines, embeddings, and vector databases for automated analysis, ranking, classification, and recommendations.
  • Engineered scalable data workflows for ingestion, enrichment, deduplication, scoring, metadata extraction, semantic matching, and delivery of structured outputs into dashboards.
  • Designed integration-ready backend patterns aligned with Salesforce, AWS, and Snowflake architectures — REST API exchange, async processing, structured data contracts, and warehouse-ready models.
  • Applied AWS serverless concepts (Lambda-style processing, API Gateway patterns, S3 storage, IAM access control, CloudWatch logging) and event-driven design for scalable AI delivery.
  • Stood up observability & evaluation — structured logs, RAG traceability, prompt evaluation, similarity metrics, and reproducible agent outputs.
FastAPINext.jsLangGraphCrewAIMCPPineconeRAGPostgreSQL
Smart Structures LLC
U.S. Army Corps of Engineers
AI Full-Stack Dev / Data Engineer
May 2025 — Aug 2025
Florida, USA
  • Designed and developed AI-enabled backend services and RESTful APIs with Python, FastAPI, LangChain, LangGraph, and GPT-4 for natural-language querying, analytics automation, and stakeholder reporting.
  • Built Generative AI, RAG, and multi-agent workflows using CrewAI, embeddings, vector databases, and prompt orchestration to automate document analysis, water-balance reporting, and operational insight generation.
  • Developed cloud-native data pipelines on GCP BigQuery, Cloud Storage, Pub/Sub, and Composer to automate ingestion, transformation, validation, and reporting of sensor and operational datasets.
  • Created anomaly-detection and engineering-diagnostics workflows with NumPy, SciPy, scikit-learn, and statistical modeling to support QA scoring, technical analysis, and decision support.
  • Implemented secure backend access controls using JWT authentication, PostgreSQL, device-bound licensing, and role-based logic for controlled desktop and internal application usage.
  • Built mobile and desktop data tools with React Native (Expo), Tkinter, PyInstaller, and cx_Oracle for field data collection, offline processing, and operational reporting.
FastAPILangChainGPT-4BigQueryPub/Subscikit-learnReact Native
South Florida Water
Management District
Research ML Engineer
Jan 2024 — Apr 2025
Florida, USA
  • Applied statistical and ML techniques — Z-Score, IQR, Isolation Forest, One-Class SVM, LOF — to detect anomalies in streaming water datasets, significantly improving QA/QC workflows and labeling accuracy.
  • Built cloud-native data pipelines on AWS using S3, Lambda, API Gateway, IAM, CloudWatch, and EventBridge to automate ingestion, anomaly evaluation, feature extraction, and reporting for large-scale hydrology datasets.
  • Developed and deployed a Generative AI chatbot with LangChain, OpenAI GPT-4, and RAG for automated documentation querying, scientific prompt generation, and real-time answer synthesis.
  • Created EDA dashboards and utilities with Matplotlib, Seaborn, SQL, real-time filtering, and interactive outlier highlighting for dynamic analysis of time-series trends and hydrologic patterns.
  • Implemented secure RBAC and JWT authentication with FastAPI, ensuring controlled access, API-level security, and data integrity across analytical applications.
  • Integrated unit testing, structured logging, real-time alerting, and CloudWatch-style observability into QA pipelines, maintaining CI/CD readiness and system reliability.
  • Collaborated closely with hydrologists, researchers, and data scientists to turn analytical needs into deployable AI, backend, and cloud-based tools.
scikit-learnFastAPIRAGAWS LambdaEventBridgeStreamlitPower BI
Deloitte USI
Gilead Sciences · State of Colorado
Data Analyst
Jan 2022 — Dec 2023
India / USA
  • Designed and optimized scalable ETL/ELT pipelines using Python, PySpark, SQL, Databricks, BigQuery, SQL Server, and AWS S3 data-lake patterns for enterprise analytics and high-volume public-sector workflows.
  • Built automated ingestion frameworks integrating REST APIs, Salesforce APIs, SOQL, flat files, AWS S3, and GCP Cloud Storage for reliable data movement between Salesforce CRM, cloud platforms, and analytics systems.
  • Developed Salesforce Public Sector workflows using OmniStudio/Vlocity, OmniScripts, DataRaptors, Integration Procedures, and FlexCards to support digital intake, eligibility-style processes, and case workflows.
  • Built event-driven Salesforce integrations with Platform Events, API callbacks, async processing, and middleware-style orchestration to improve data synchronization across cloud pipelines.
  • Implemented real-time and batch pipelines with Kafka, Spark Structured Streaming, Spark SQL, Hadoop, and Databricks notebooks for event-driven transformations and proactive alerts.
  • Developed semantic data models, SQL transformations, and materialized views in BigQuery, SQL Server, and Snowflake-style warehouses powering Power BI and Tableau dashboards with optimized performance.
  • Created data-quality, anomaly-detection, and reconciliation scripts with Pandas, NumPy, scikit-learn, and SQL, improving data quality by 40% and strengthening audit readiness.
  • Deployed containerized ETL/ML workflows using Docker, Airflow, Cloud Composer, GCP Cloud Functions, MLflow, and Vertex AI with CI/CD practices.
PySparkDatabricksKafkaSalesforceAirflowVertex AITableauSnowflake
03 — VENTURES

Things I've founded & built.

A portfolio of AI-native products across media, careers, education, cloud, and developer tooling.

FinalDraft AI

Founder

AI-native drafting and document-generation engine that turns rough inputs into polished, structured final drafts — with editing, formatting, and review workflows.

Writing · GenAI

ClonA AI

Founding Eng

AI dubbing & lip-sync workflows combining video pipelines, voice processing, model orchestration, and GPU-backed inference for creative-studio media use cases.

Media · Multimodal

Knuckler AI

Founder

AI career platform — resume intelligence, job matching, document parsing, scoring, and outreach, with Salesforce + Apollo CRM-enriched candidate and company data.

CareerTech

SaSa Launch AI

Co-Founder

AI SaaS built on cloud agents — multi-agent orchestration, architecture and product decisions, and rapid 0→1 MVP execution with founder-style ownership.

Cloud · Agents

GradeUS AI

Co-Founder

AI grading platform for instruction-aligned assessment, rubric-based feedback, and hallucination-aware evaluation, with an OCR-to-assessment multi-model vision pipeline.

EdTech · Agents

Visra

AI SaaS Growth

AI dubbing SaaS across 12+ languages — GPU-accelerated video processing with face restoration, anti-tearing, and multi-language TTS integration.

Media · Multimodal

Sakura AI

Founder

AI-native workflow automation with Stripe billing, usage metering, multi-tenant architecture, and a developer API portal built in from day one.

SaaS · Infra

BoxOps AI

AI Engineer

AI-enabled operational workflows for film distribution — activity tracking, weekly reporting, representative submissions, and invoice automation.

Ops · Automation
04 — SELECTED PROJECTS

Engineering, shipped & shown.

A sample of open and award-winning work — most production work lives in private repos.

AI-Enhanced Big Data
Pipeline Framework

PySpark + Airflow + Kafka pipeline for intelligent ETL — streaming ingestion, anomaly-detection models, and MLflow tracking. Model-driven rules detect schema drift, outliers, and pipeline failures for reliable high-volume processing.

PySparkAirflowKafkaMLflow

Healthcare Intelligence
Pipeline

Python + FastAPI + Streamlit pipeline integrating ETL, ML models, and CrewAI multi-agent workflows. Readmission-risk prediction (Random Forest, Gradient Boosting) at 85% accuracy with real-time patient processing in an interactive dashboard.

FastAPICrewAIStreamlitscikit-learn

AI-ML-Analytics
Assistant

LangChain + OpenAI + scikit-learn analytics platform. Users connect datasets, ask questions in plain English, and trigger ML workflows — fusing LLM reasoning with classic ML for classification, regression, and explainable insights.

LangChainOpenAINL AnalyticsML Deploy

Maternal Health in 3D
AI Medical Imaging

★ NIH ALL OF US — WINNER

3D maternal-health monitoring with OpenCV, Open3D & Intel RealSense for gait analysis, fall detection, and posture tracking — converting depth signals into spatial insights for pregnancy outcome prediction.

OpenCVOpen3DRealSenseComputer Vision

IoT Table-Top
CNC Machine

⬡ PATENT · 352285-001

IoT-based table-top CNC application on Raspberry Pi — +40% operational efficiency, −30% setup time. Led embedded device programming and real-time software; awarded a design patent.

Raspberry PiEmbeddedIoTReal-time

ATV Quad Bike
Design Challenge

▲ NATIONAL CHAMPIONS

Team Captain & Chief Technical Designer — led the team to the QBDC Season 5 National Championship, using Python, SQL & ML to evaluate performance data and improve design accuracy ~25%.

PythonPower BIMLLeadership
05 — STACK

The toolkit.

Deep, hands-on across the modern AI and data stack — from prompt, to pipeline, to production deployment.

GenAI & Agents
LangChainLangGraphCrewAIMCP ServersMulti-Agent SystemsRAGEmbeddingsPrompt EngineeringFunction / Tool UseStructured OutputsFine-tuningEvaluationObservabilityGPT-4 / 5ClaudeGeminiLLaMA / QwenDeepSeekHugging Face
Languages & Backend
PythonTypeScriptSQLRCFastAPINext.jsReactNode.jsREST APIsWebhooksAsync ProcessingMicroservicesPydanticJWT · RBACVercel AI SDK
Vector & Retrieval
PineconeFAISSChromaDBpgvectorWeaviateSemantic SearchHybrid SearchReranking
Data & ML
PySparkDatabricksKafkaAirflowSpark StreamingHadoop · HDFSPandas · NumPySciPyTensorFlowPyTorchscikit-learnXGBoostMLflowAnomaly Detection
Cloud & Platforms
AWS Lambda · S3API GatewayEventBridgeSQS · SNS · KinesisStep FunctionsBedrockSageMakerGCP BigQueryVertex AIPub/SubCloud ComposerDataflowSnowflake · SnowparkOracle Cloud
Voice, Video & Creative
ElevenLabsLiveKitDeepgramCartesiaMulti-language TTSLip-Sync / DubbingGPU InferenceFFmpegThree.js · GLSLOpenCV · Open3D
Data Platforms & Storage
PostgreSQLSupabaseNeonDrizzle ORMSQL Servercx_OracleSalesforceOmniStudio / Vlocity
Ops, Tooling & BI
DockerKubernetesVercelGitHub ActionsCI/CDStripeOCR · NLPMathpix · Google VisionPower BITableauPlotlyCursor · Windsurf
06 — RECOGNITION

Proof points.

NIH AI Hackathon Winner

1st place — NIH All of Us AI Hackathon for 3D AI medical imaging.

Granted Patent

IoT-based table-top CNC machine · Application No. 352285-001.

National Champions

Team Captain — QBDC Season 5 Quad Bike Design Challenge.

M.S. Data Science

Florida Atlantic University · 3.89 / 4.0 GPA · Research Assistant.

Certifications
Google Cloud Professional Data Engineer
Salesforce Agentforce Specialist In progress
Microsoft Power BI Data Analyst Associate
Deloitte AI Academy Certified
Oracle Cloud AI Foundations
FAU Big Data Engineer · CITI RCR
07 — WORK AUTHORIZATION

Ready to work in the U.S.

Clear, secured authorization — the hardest step is already cleared.

Authorized to work in the U.S.

No long-term
visa uncertainty.

I'm currently authorized to work and my H-1B has already been selected in the lottery — so the path to long-term authorization is secured, not speculative.

Current status — F-1

On an F-1 student visa with active employment authorization (OPT). Eligible to start work immediately.

Work-authorized now

H-1B — selected in the lottery

Cap-subject H-1B petition selected and in process, securing continued long-term employment authorization.

Selected ✓

What this means for an employer

I can begin work now on existing authorization, and the H-1B selection removes the biggest sponsorship risk for a long-term hire.

Let's talk

Building something ambitious?

I'm open to founding-engineer, AI-engineer, and product roles in the U.S. — and always happy to talk shop on agents, RAG, and 0→1 product.

saiprasadmuppala2602@gmail.com +1 415-470-7299 github.com/msaiprasad1