Available for research collaboration
GenAI platforms, agentic workflows, research, and open source

Goutam Adwant

Principal Engineer & Independent Researcher

Principal Engineer in an S&P 500 enterprise focused on GenAI and agentic solutions using AWS services, Python, LangGraph, MCP, LLMs, and in-house AI infrastructure.

Years building distributed systems
13+
Research papers and datasets
8
Published open-source packages
12
AI-focused papers reviewed
65+

Profile Highlights

Research, open source, and current engineering leadership

Organized around public work, current production AI platform responsibility, and the systems that connect research with shipping software.

S&P 500 enterprise

Current Principal Engineering Role

Building production GenAI and agentic systems using AWS services, Python, LangGraph, MCP, LLMs, and in-house AI infrastructure, while shaping secure, scalable platform capabilities inside a large public enterprise.

  • Chatbots, agentic workflows, and AI-powered product capabilities
  • AWS services, Python, LangGraph, MCP, LLMs, and in-house AI infrastructure
  • Architecture and infrastructure stacks for production AI platforms

Public manuscripts

Research Papers & Active Work

Research manuscripts and datasets across generative AI policy, RAG evaluation, hybrid cloud-edge intelligence, model editing, and knowledge agents.

  • 8 research works
  • TechRxiv preprints
  • Zenodo dataset
  • Direct paper links

Public engineering output

Open Source Systems

Published libraries for MCP security, RAG evaluation, LLM fleet routing, performance linting, tracing, schemas, and prompt contracts.

  • 11 PyPI packages
  • PromptSpec-J on Maven
  • 47 upstream OSS contributions
  • Technical blog

Verified reviewer activity

Peer Review Service

Reviewed 65+ AI-focused manuscripts across IEEE, ACM, and Springer venues, with additional reviews in progress across journals and conferences with selective technical review standards.

  • IEEE Access and IT Professional
  • IEEE TPAMI and Neural Computing and Applications
  • ACM AI and Agentic Systems, ACM TOSEM, IEEE TNSE

Research & Publications

Research papers and active work

Publications and datasets across generative-AI policy, retrieval-augmented generation, hybrid cloud-edge intelligence, model editing, and knowledge-agent architectures.

8

Research works

7

Current preprints

Active journal and conference targets
IEEE SoftwareIEEE Communications LettersACM GoodITSUSTAINIEEE ISAIA

Academic Service & Peer Review

Invited reviewer across premier AI venues

Serving as an invited reviewer for top-tier IEEE, ACM, and Springer journals and conferences, with 65+ AI-focused manuscript reviews completed and additional reviews in progress.

65+

AI-focused reviews

9

Venues served

3

Publishers

Professional Memberships
IEEE

IEEE Access

Invited Reviewer

IEEE

IEEE Transactions on Pattern Analysis and Machine Intelligence

Invited Reviewer

ACM

ACM Conference on AI and Agentic Systems

Invited Reviewer

IEEE

IEEE Transactions on Mobile Computing

Invited Reviewer

ACM

ACM Transactions on Software Engineering and Methodology

Invited Reviewer

Springer

Neural Computing and Applications

Invited Reviewer

IEEE

IEEE Open Journal of the Computer Society

Invited Reviewer

IEEE

IEEE Transactions on Network Science and Engineering

Invited Reviewer

IEEE

IT Professional

Invited Reviewer

Reviewer Certifications

Introduction to Peer Review

Web of Science Academy

View credential

Certified Peer Reviewer

Elsevier BV (Netherlands)

View credential

ACM Certified Peer Reviewer

Association for Computing Machinery

View credential

Serving as invited reviewer across IEEE, ACM, Springer

Engineering Open Source

Published libraries and developer tools

12 published packages

12 published packages

bigocheck

bigocheck

Zero-dependency, AI-assisted Big-O complexity checker. Static analysis + empirical benchmarking for Python.

PythonPyPIcomplexitybig-o

kvfleet

kvfleet

Production-grade, KV-cache-aware intelligent routing for self-hosted and hybrid LLM fleets.

PythonPyPIgpu-routinginference-routing

mcp-egress-guard

mcp-egress-guard

Local-first MCP reverse proxy that blocks sensitive, destructive, or policy-violating tool calls before execution using deterministic rules, DLP matchers, and AST-based intent analysis.

PythonPyPIagent firewalldata loss prevention

mcp-pool

mcp-pool

Async connection pool for Model Context Protocol (MCP) client sessions — keep sessions warm, reuse across requests, auto-reconnect on failure.

PythonPyPIagentasync

mcp-shield-pii

mcp-shield-pii

Intercepting gateway proxy for MCP clients/servers — real-time PII redaction with regex, NLP, and optional subinterpreter concurrency

PythonPyPIgdprhipaa

pdperf

pdperf

A static performance linter that detects slow Pandas anti-patterns before they reach production.

PythonPyPIpandasperformance

pydanticforge

pydanticforge

Infer robust Pydantic v2 models from messy, evolving JSON streams

PythonPyPIclicode-generation

schemaglow

schemaglow

Human-friendly schema diff and contract drift detection for CSV, JSON, JSONL, Parquet, OpenAPI, Avro, and protobuf.

PythonPyPIavrocli

toolcallcheck

toolcallcheck

Deterministic Python testing for tool-using agents. Mock MCP tools, assert exact tool calls and trajectories, verify headers, and run offline in CI.

PythonPyPIagent-testingai-agents

tracemap

tracemap

Modern traceroute visualizer for the terminal with TUI, interactive HTML maps, and ASN/GeoIP lookups.

PythonPyPIasciidebugging

vectormigrate

vectormigrate

Python-first tooling for safe embedding-model migration across vector retrieval systems.

PythonPyPIai-infrastructureembeddings

promptspecj

PromptSpec-J

A suite of Java libraries for LLM prompt specifications. Includes sub-modules: promptspec-model, promptspec-parser, promptspec-validator, promptspec-runtime, promptspec-codegen-java, promptspec-spring-ai-adapter, promptspec-junit5, and promptspec-maven-plugin.

JavaMavenPrompt Engineering

External Open Source Contributions

Upstream framework fixes

47 upstream contributions

apache/seatunnel · #10987 · Merged Jul 25, 2026

Use checkpoint offset for timestamp startup restore

Fixed SeaTunnel MySQL CDC checkpoint restore behavior for timestamp startup mode so restored jobs resume from the saved binlog position.

Merged upstream5 files+133 / -3MySQL CDC restore
Issue

For startup.mode = timestamp, SeaTunnel could recompute the original configured timestamp after a checkpoint restore instead of honoring the concrete binlog file and position already persisted by the reader. In long-running jobs, that could move recovery back to an old timestamp anchor and fail when older binlogs were purged or corrupted.

Fix

Added a timestamp-only offset guard, reused restored binlog offsets directly during startup offset resolution, skipped the timestamp filter for restored concrete offsets, and pinned the behavior with focused MySQL CDC unit tests.

Technical value
  • Improves reliability for long-running MySQL CDC pipelines that use timestamp startup and checkpoint recovery.
  • Prevents restored jobs from rewinding to the original timestamp when a checkpoint already contains a concrete binlog position.
  • Avoids unnecessary timestamp-to-binlog lookup and timestamp filtering on the restore path.
Validation
  • Added unit coverage for timestamp-only bootstrap offsets, restored binlog offsets, and binlog timestamp-filter decisions.
  • Verified with the connector Maven package command documented in the PR.
  • Merged after maintainer approval on July 25, 2026, closing the linked MySQL CDC recovery issue.
BinlogOffset state modelstartup offset resolutionbinlog timestamp filteringcheckpoint restore semanticsMySQL CDC regression tests
JavaSeaTunnelMySQL CDCCheckpoint restore

Experience Timeline

Leadership arc across systems engineering and AI research

A concise progression through graduate systems work, platform engineering, principal engineering responsibility, and independent research output.

  1. 2014 - 2016

    M.S. in Software Engineering

    San Jose State University

    Graduate foundation in software architecture, scalable systems, and disciplined engineering delivery.

    Software architectureDistributed systemsResearch
  2. 2015 - 2017

    Big Data and Platform Engineering

    Hewlett Packard Enterprise

    Built data-intensive platform capabilities and production pipelines across enterprise engineering environments.

    JavaPythonData pipelinesPlatform
  3. 2017 - Present

    Principal Engineer

    S&P 500 enterprise

    Builds GenAI and agentic solutions using AWS services, Python, LangGraph, MCP, LLMs, and in-house AI infrastructure. Designs scalable, secure infrastructure stacks for production AI platforms.

    AWS servicesLangGraphMCPPython
  4. Current

    Independent Researcher

    Applied AI and open-source systems

    Bridges production AI architecture with LLM evaluation, RAG reliability, cloud-edge AI behavior, and public developer tooling.

    LLMsRAGAWSOpen source

Technical Depth

Engineering stack and platform depth

Hands-on architecture depth across Java, Python, Spring Boot, Docker, Terraform, CI/CD, applied AI infrastructure, and data-intensive microservices.

Cloud Architecture

AWS-first distributed systems with pragmatic reliability boundaries.

AWSTerraformDockerCI/CDObservability

Backend Systems

Data-intensive services built for maintainability and operational clarity.

JavaPythonSpring BootMicroservicesREST

Applied AI

Production-ready GenAI, agentic workflow, RAG, and evaluation pipelines with clear measurement and traceability.

LLMsRAGLangGraphMCPAgentic workflows

Developer Platforms

Tooling that moves quality gates earlier into local development and CI.

ASTMCPPyPIMavenAutomation

From the Blog

Technical notes from the build path

Long-form writing on AI infrastructure, distributed systems, and the engineering trade-offs behind public libraries.

View all posts

Contact

Let's build something meaningful

Whether it's a research collaboration, open source contribution, or technical advisory — I'm always open to conversations that push the boundaries of AI and distributed systems.

Based in Livermore, California.

Pacific Time (PT) · Usually within 24 hours

workwithgoutam@gmail.com
Use email app