99.7%
Deployment reliability
Turned recurring release failures into a steadier deployment path across Jenkins, Docker, and AWS EC2.
Parvez Shaik - SDE / SRE / Cloud Engineer
Systems that scale.
Software engineer with an MS in Computer Science from Indiana University and 4 years across backend, cloud reliability, DevOps automation, and AI retrieval systems.
Open to SRE, DevOps, cloud, and backend roles across the U.S.
01 - Impact
Measured engineering impact.
420 -> 295 ms
PostgreSQL latency reduction
Found the slow path, tightened the database layer, and made production reads noticeably faster.
35 min
Release time saved per sprint
Moved release checks earlier so teams spent less time chasing avoidable environment issues.
23%
Review effort reduced
Added agent-assisted evaluation to an AI content pipeline, helping reviewers move faster without losing oversight.
02 - Work
Featured projects.
Featured project
Agentive Med
A healthcare-focused AI assistant designed around guarded orchestration: triage routes the question, domain agents isolate medical and nutrition context, and a verifier checks answers against retrieved clinical sources.
PythonAG2AutoGenFAISSPubMed RAGMulti-Agent Systems
Results
Supervisor-worker orchestration
PubMed-backed retrieval
Verifier agent guardrail
Featured project
Distributed Vector Database for AI Applications
A retrieval system built around fast semantic search, distributed node communication, and storage designed for large LLM workloads.
PythongRPCRocksDBLlamaIndexAWS S3GraphRAG
Results
140M+ embeddings
93 ms average query latency
Context-aware search at scale
03 - Experience
Professional experience.
01
Jun 2025 - Aug 2025
Software Developer Intern
MyEdMaster
GraphRAGAutoGenAWS EC2RedisMySQL
Helped turn a manual content review process into an agent-assisted retrieval workflow with faster evaluation and stronger relevance.
- Introduced agent-assisted review into a Python content pipeline, cutting evaluation effort by 23% while keeping the workflow accountable.
- Strengthened retrieval quality with LlamaIndex, SpaCy, and GraphRAG, helping generate 2,800+ assessment items as relevance rose from 72% to 83%.
- Stabilized the Flask deployment path on AWS EC2 with Docker and Apache, removing repeated backend friction and improving evaluation speed by 18%.
02
Sep 2022 - Apr 2024
Software Development Engineer II
Cognizant
Node.jsPostgreSQLJenkinsDockerAWS EC2
Worked close to production systems where latency, failed builds, and release reliability had direct team impact.
- Traced bottlenecks through Node.js services and PostgreSQL, reducing production read latency from 420 ms to 295 ms.
- Used logs, runtime checks, and release patterns to push Jenkins, Docker, and AWS EC2 deployment reliability to 99.7%.
- Shifted configuration and smoke-test checks earlier in the pipeline, saving 35 minutes per sprint and reducing deployment failures by 45%.
- Kept sprint releases moving by translating failed builds and environment issues into clear fixes across development, QA, and DevOps.
03
Aug 2021 - Sep 2022
Software Development Engineer
Cognizant
ReactGraphQLDjango APIsDockerJenkins
Balanced internal product work with the infrastructure needed to make environments faster to set up and easier to trust.
- Built React analytics views backed by GraphQL and Django APIs for near real-time internal reporting.
- Reworked fetch and render behavior so dashboard updates felt faster, improving rendering speed by 21.6%.
- Containerized backend environments on AWS EC2 with Docker and Jenkins, cutting setup time from 82 minutes to 39 minutes.
04
Jan 2021 - Jul 2021
Software Development Intern
Widhya
TensorFlowSpaCyPandasScikit-learn
Built early discipline around experiments: cleaner preprocessing, repeatable evaluation, and measurable model improvement.
- Refined tokenization and model settings for an RNN text classifier, raising validation accuracy from 88.4% to 93.2%.
- Automated preprocessing and evaluation scripts with Pandas and Scikit-learn, reducing repetitive testing time by 27%.
04 - Strengths
Engineering focus.
Backend Systems
Service layers that stay readable under load: APIs, caching, auth, and data paths teams can operate.
Cloud Reliability
Release paths with fewer surprises: logs, smoke tests, runtime checks, and cleaner handoffs.
AI Retrieval
LLM workflows grounded by retrieval, evaluation loops, and the data structures behind useful answers.
05 - Skills
Technical toolkit.
Languages
PythonJavaScriptTypeScriptJavaJ2EEJDBC
Backend and APIs
Node.jsExpress.jsFastAPISpringHibernateRESTGraphQLgRPC
Data Stores
PostgreSQLMySQLRedisMongoDBRocksDBFAISS
Cloud and DevOps
AWSAzureDockerKubernetesJenkinsGitHub ActionsTerraformLinux
AI and Retrieval
RAGGraphRAGLlamaIndexAutoGenSpaCyLLM Evaluation
Reliability
ObservabilityLogsIncident ResponseSmoke TestsLatency OptimizationRelease Support