Reethika ValluruLet's connect ↗AI engineer · Story collector · Curious human
I'm Reethika—an AI/ML and automation engineer turning complex problems into practical products, while documenting the places, lessons, and ideas that shape my journey.
Between code & the open road
I'm a software engineer with six years of experience across AI/ML, Python development, test automation, and software validation.
My work sits where reliable engineering meets applied intelligence: building automation frameworks, experimenting with retrieval and language models, and translating prototypes into useful workflows.
Outside work, I travel to notice the details—new cities, long roads, local food, mountain light—and write down what those experiences teach me.
A selection of AI, automation, and engineering work focused on solving real operational problems.
Applied AI · NLP
An end-to-end AI defect-management workflow—from intelligent triage and ownership mapping to root-cause analysis and suggested code fixes.
An AI-powered microstructure characterization platform that segments ferrite, pearlite, martensite, and inclusions; measures grain size and phase fractions; detects cracks, pores, and defects; and generates automated metallographic inspection reports.
Engineering · Automation
Python automation for storage-controller configuration, firmware upgrades, and repeatable system validation in complex server environments.
A seven-day multimodal agentic-AI platform implementing vision-language ingestion, hybrid GraphRAG, episodic memory, MCP-native tool execution, multi-agent orchestration, constrained decoding, and end-to-end LLM observability.
LLMOps · Evaluation Engineering
A continuous LLM-evaluation control plane implementing dataset versioning, prompt regression, RAGAS metrics, trajectory grading, calibrated LLM-as-judge ensembles, adversarial red teaming, drift detection, and release-gated model promotion.
Streaming ML · Anomaly Intelligence
A low-latency anomaly-intelligence pipeline combining Kafka event ingestion, online feature computation, temporal transformers, isolation forests, Bayesian change-point detection, SHAP attribution, concept-drift monitoring, and automated incident correlation.
Edge AI · Privacy-Preserving ML
A federated computer-vision training substrate with differential privacy, secure aggregation, non-IID client sampling, LoRA parameter-efficient fine-tuning, quantization-aware training, ONNX graph optimization, and hardware-adaptive edge inference.
Practical AI tutorials, honest career reflections, and travel stories from life across India and Europe.
7 min readIntense beginnings, unfinished endings, social masking, restless attention—and the slow work of replacing shame with curiosity.
Read the story ↗
7 min readHow a room full of nervous strangers becomes a table full of laughter—and why playing together feels so much better than another weekend spent scrolling.
Read the story ↗
6 min readWhat learning German has taught me about grammar, patience, mistakes, and the very different ways we learn languages as children and adults.
Read the story ↗Developed tools for configuration, firmware workflows, and reliable validation.
Built and maintained automated tests with Pytest, Selenium, and Jasmine.
Applied NLP, embeddings, retrieval, and experiment tracking to engineering workflows.
Open to thoughtful conversations
Let's talk about AI, engineering, travel—or the interesting places where they meet.
Say hello ↗hello@reethika.in