AI/ML & Software Engineer
Applied AI · Data Systems · Backend Engineering · Reliability & Cloud

I am a Computer Science engineer from Hyderabad who enjoys building practical, testable software at the intersection of machine learning, data engineering and backend systems.

My recent work includes decision-support platforms, scientific data pipelines, secure RAG applications, cloud-native APIs, observability tooling and production-minded AI systems.

• 🎓 B.Tech in Computer Science and Engineering, 2022–2026
• 📈 CGPA: 8.94/10
• 🏅 GATE 2026: All India Rank 3,863
• 💼 Academic Intern: iSpace Software Solutions
• 🛰️ Former AI Research Intern: NRSC, ISRO
• 📍 Hyderabad, India
• 📧 sai.kumar.garlapati.cs@gmail.com
• 🔎 Open to entry-level opportunities in AI/ML, software engineering, backend, data and cloud

Explainable ML models, retrieval-augmented generation, LLM-backed workflows, recommendation systems and intelligent decision-support applications.

Automated data discovery, ingestion, validation, incremental synchronization, structured storage, analytics and API-based access.

FastAPI and Java services, relational databases, secure API design, Docker, CI/CD, observability, Kubernetes and OpenShift-ready applications.

Evidence-grounded outputs, deterministic fallbacks, schema validation, bounded failure modes, responsible-AI controls and human-reviewable decisions.

A customer-value, propensity-modeling and marketing-investment optimization platform built using synthetic data.

Highlights

• Customer conversion propensity and contribution-value modeling
• Auditable campaign targeting and budget allocation
• Experiment analysis with confidence intervals and sample-ratio checks
• Bronze, Silver and Gold data architecture
• FastAPI serving layer and Streamlit dashboard
• CSV and Excel exports
• Docker Compose and automated quality checks

Stack: Python, XGBoost, scikit-learn, FastAPI, Streamlit, Pandas, PyArrow, PostgreSQL, Docker

Status: In progress, with a working MVP implemented. Current work focuses on production hardening, deployment and optimization improvements.

An observable reliability-engineering lab for investigating deterministic service failures and generating evidence-grounded AI analysis.

Highlights

• Deterministic success, failure and latency simulations
• Request-ID propagation and structured JSON logging
• OpenTelemetry and Azure Application Insights readiness
• Gemini, Groq and deterministic local-provider fallback
• Pydantic-validated structured AI outputs
• Bounded telemetry, safe diagnostics and responsible-AI controls
• Extensive automated testing and CI

Stack: Python, FastAPI, Pydantic, OpenTelemetry, Azure Monitor, Gemini, Groq, Docker

An end-to-end platform for discovering, ingesting, storing and analysing public India WRIS river-discharge data.

Highlights

• More than 650,000 hydrological measurements
• Data collection across 188 monitoring stations
• Automated station discovery and parallel ingestion
• Incremental synchronization for newly available records
• Structured storage and REST API access
• Streamlit analytics dashboard
• Natural-language-to-SQL querying

Stack: Python, SQLite, FastAPI, Streamlit, SQL, Groq, concurrent data pipelines

This is an independent educational portfolio implementation using public data. It does not represent or expose confidential NRSC or ISRO systems.

A production-minded FastAPI service demonstrating application health semantics, observability and secure OpenShift-ready deployment.

Highlights

• Separate liveness and readiness endpoints
• Concurrent dependency checks with bounded timeouts
• RFC 7807-style error responses
• Prometheus metrics with controlled label cardinality
• Structured logs and sanitized request identifiers
• Red Hat UBI-based non-root container
• Kustomize development and production overlays
• CI, dependency auditing and container-security validation

Stack: Python, FastAPI, Prometheus, Docker, OpenShift, Kubernetes, Kustomize, GitHub Actions

A Java 21 backend for tenant-isolated device inventory and safe TCP-reachability monitoring.

Highlights

• Versioned REST APIs using Jakarta Servlets and Jetty
• Tenant isolation across queries and mutations
• MySQL persistence using JDBC, HikariCP and Flyway
• Concurrent polling using ExecutorService and CompletableFuture
• Duplicate-poll suppression and bounded socket timeouts
• Docker Compose and Kubernetes deployment assets
• Health checks, metrics and OpenAPI documentation

Stack: Java 21, Jetty, MySQL, JDBC, Flyway, Docker, Kubernetes, Maven

An enterprise-style internal knowledge assistant with authorization enforced before retrieval and generation.

Highlights

• JWT-based authentication
• Role-based document access
• Semantic retrieval with metadata filtering
• Retrieval-augmented generation with source attribution
• Provider-independent LLM integration
• Streamlit user interface
• Role and misuse validation tests

Stack: Python, FastAPI, Chroma, Sentence Transformers, JWT, Streamlit, SQLite

April 2026 – Present

• Developing Python and JavaScript components for software-modernization workflows
• Working with browser automation, API integration and legacy-system analysis
• Improving modularity, error handling and diagnostic logging
• Investigating API and network behaviour using structured logs and HAR captures
• Preparing implementation documentation and maintainable project workflows

January 2026 – March 2026

• Worked with more than 650,000 hydrological observations
• Built automated data-discovery, ingestion and validation workflows
• Implemented incremental synchronization and structured error recovery
• Developed database-ready storage and REST access components
• Explored Physics-Informed Neural Networks for scientific machine learning

December 2025 – January 2026

• Developed a secure internal knowledge assistant
• Implemented retrieval-augmented generation and role-based access control
• Added JWT authentication, metadata-filtered retrieval and source attribution
• Structured the system into modular backend, retrieval and frontend components

March 2025 – June 2025

• Built machine-learning workflows using Python
• Performed data preprocessing, model development and evaluation
• Worked on practical AI and analytics use cases

Python Java SQL JavaScript C Bash Dart

PyTorch TensorFlow scikit-learn XGBoost SHAP Transformers OpenCV RAG LLM APIs

FastAPI Pydantic Flask PostgreSQL MySQL SQLite Pandas PyArrow Streamlit

Docker GitHub Actions Linux OpenTelemetry Prometheus Kubernetes OpenShift Microsoft Azure

Automated Testing Type Checking Dependency Auditing Structured Logging API Design Security Reviews CI/CD

• 🏅 GATE 2026 — All India Rank 3,863
• 🎓 B.Tech CSE — 8.94/10 CGPA
• 🛰️ AI Research Internship at NRSC, ISRO
• 📊 Built and documented multiple production-minded AI and backend portfolio systems
• 🧪 Achieved strong automated-test coverage across recent backend and ML projects

• Oracle OCI Generative AI Professional
• Oracle Cloud Infrastructure AI Foundations
• Oracle Vector Search Professional
• Google AI Essentials
• Microsoft Azure AI
• ISRO Geodata Processing using Python and Machine Learning
• Harvard CS50P
• Harvard CS50SQL
• Harvard Introduction to Cybersecurity
• Cisco Introduction to Cybersecurity
• Postman API Student Expert

Currently expanding my knowledge in:

• Agentic AI engineering
• MLOps and production ML
• Scientific machine learning
• Cloud-native systems
• Cybersecurity
• Quantum computing

• Make technical claims verifiable through code, tests and documentation.
• State limitations clearly instead of hiding them behind “production-ready” labels.
• Keep AI outputs evidence-grounded and human-reviewable.
• Prefer reproducible workflows and deterministic behaviour.
• Design secure defaults before adding convenience.
• Build systems that fail safely and explainably.

Building useful systems, documenting the trade-offs and learning in public.