Looking for a job

v Saikiran

@vsaikiran

Message

Technical Leader Researching and building scalable AI pipelines for logs, documents, and LLMs.

India

Message

What I'm looking for

I’m looking to build production-grade ML/LLM systems—log and document intelligence with RAG, efficient retrieval, and scalable data pipelines—where strong engineering, measurable performance, and real-world deployment matter.

I lead development of Splunk’s log foundation encoder, delivering a general-purpose embedding backbone for log pruning, anomaly detection, and log reasoning. I built an end-to-end pipeline that distilled 10B raw log events into a curated 200M-event training corpus, pretrained on AWS SageMaker, and introduced Field-Aware Masking and a weak-decoder pretraining approach that outperformed evaluated baselines.

I also ideated and shipped PwnyGPT, Splunk’s internal chatbot engine powered by LLaMA 3.1 70B with Retrieval-Augmented Generation over the documentation corpus. Previously, at Brane Enterprises, I developed generalized document extraction using an LLM + deep learning architecture, optimized deployment with VLLM, and improved turnaround time by 50% through profiling and systems tuning; earlier at CISCO, I enhanced AWS auto-scaling and integrated telemetry into Azure Application Insights. I’m passionate about building reliable, production-ready AI systems—grounded in strong engineering, measured by real performance, and protected via innovation like patents.

Experience

Work history, roles, and key accomplishments

Current

Group Technical Leader I

Current

Splunk (Cisco)

Led development of a 110M-parameter log foundation encoder for Splunk AI, including an LSH-based data pipeline, field-aware masking, and a weak-decoder pretraining approach. Also ideated and built PwnyGPT, Splunk’s internal RAG chatbot over Splunk’s documentation, using multi-stage retrieval and microservices.

Foundation Models PyTorch AWS Sagemaker Distributed Training S3 Transformers LSH Based Deduplication Masked Language Modeling Field Aware Masking Weak Decoder Pretraining LLaMA RAG Weaviate DynamoDB fastAPI Kubernetes Microservices BM25 Unit Testing

Software Developer (Security BU)

Cisco

Jul 2019 - Feb 2021 (1 year 7 months)

Enhanced AWS auto-scaling for the Cisco ASAv firewall using Python-based Lambda metrics integration. Integrated security firewall metrics from a vFTD device into Azure Application Insights using Telegraf via a Go-language plugin.

Python AWS Lambda EC2 S3 EFS CloudFormation ECR Telegraf Go Azure App Insights Security Metrics IaaS

Summer Internship Project

University of Nebraska-Lincoln

Implemented a skin lesion image classification architecture for 7 classes and studied Variational Autoencoders (VAEs) and GANs to generate new images with fewer training examples.

Skin Lesion Image Classification Variational Auto Encoders Gan Image Generation Deep Learning

Summer Internship Project

IISc Bangalore

Implemented a Particle Swarm Optimization (PSO) algorithm for constraint-less optimization of curve and developed an approach for multi-constraint curve optimization.

Particle Swarm Optimization (PSO)Optimization Curve Optimization Constraint Optimization

Machine Learning Scientist II

Brane Enterprises

Built a generalized document extraction engine for extracting key-value pairs across finance and healthcare, including an LLM-enhanced pipeline, VLLM optimization, and microservice architecture. Led additional work on contextual PDF question answering via RAG and developed a mobile Tangram game using real-time shape detection with OpenCV.