I build production AI systems and enterprise frontend platforms. I care about reliability, observability, and shipping useful software fast. Most of my production delivery work is private by design, and the public repositories here are representative demos, reference implementations, and exploratory builds.
- ML Engineer at Accenture India (R&D)
- GCP Certified Professional Machine Learning Engineer: https://www.credly.com/badges/8c5f8591-32b6-41e7-bacf-5d12780899c3/linked_in_profile
- Hyderabad, India
- Portfolio: https://ramdragneel01.github.io/dragon-portfolio/
- LinkedIn: https://linkedin.com/in/ramprakashdhulipudi
- Medium: https://medium.com/@RamPrakashD
For the latest execution updates: Profile Review Follow-Up (May 2026)
| Signal | Current evidence |
|---|---|
| GitHub contribution consistency | 1,645 contributions in the last year |
| Public repository base | 44 public repositories |
| External OSS visibility | 1 external PR opened + 1 meaningful issue comment |
| Measured outcomes documentation | 6 of 6 flagship repositories |
| Delivery focus | Trustworthy AI, reliability engineering, and enterprise frontend platforms |
- Trustworthy RAG: evaluation beyond LLM-as-judge loops
- ML observability: drift, latency, and reliability in production
- Enterprise frontend: micro-frontends with governance and contracts
| Repository | Why it exists | Current direction |
|---|---|---|
| hallucination-lens | Measure RAG faithfulness at sentence level | Add 3 benchmark datasets and ship v1 CLI JSON output |
| context-watchdog | Guardrails for long-running LLM and agent workflows | Publish pip package and policy benchmark report with 3 recipe presets |
| agentic-research-assistant | Multi-agent research with traceable orchestration | Add deterministic eval suite with citation-faithfulness and cost-tier metrics |
| mlops-sentinel | Monitor model behavior in production | Add alert triage dashboard, SLO thresholds, and monthly trend snapshots |
| partner-portal-microfrontends | Enterprise portal using federated React apps | Publish live mock-auth demo with badge and accessibility evidence |
| dragon-portfolio | Public case-study surface for shipped systems | Add quantified impact cards and monthly release radar updates |
| Repository | What reviewers should look for first |
|---|---|
| hallucination-lens | Faithfulness scoring approach + benchmark framing |
| context-watchdog | Policy guardrail patterns for long-running agent workflows |
| agentic-research-assistant | Multi-agent orchestration and traceability |
| mlops-sentinel | Drift and reliability monitoring loop |
| partner-portal-microfrontends | Enterprise frontend governance and contracts |
| dragon-portfolio | Public case-study index and collaboration entry point |
- Clear architecture and reproducible setup in every repository
- Automated checks (lint, test, build) on every meaningful change
- Evidence of outcomes: latency, cost, and reliability improvements
- Small, frequent releases instead of large infrequent drops
I am building AI Reliability Chaos Lab as the next flagship project to pin.
- Plan document: AI Reliability Chaos Lab - 3-week build plan
- Target outcome: a production-grade reliability chaos platform for LLM and agent systems with reproducible failure scenarios, mitigation policies, and release-grade evidence.
- Pin-ready target: semantic release
v1.0.0after week 3 validation and documentation hardening.
- External contribution proof: increase to at least two merged external PRs and five meaningful issue comments, then publish direct evidence links using tracker and evidence template.
- Replace remaining placeholder metrics with measured numbers in all flagship README results blocks.
- Keep release cadence consistent with one semantic tag per flagship repository each month.
- Add one concrete screenshot or short demo GIF to each flagship repository README.
- Reduce container CVE warnings in flagship repos and document residual risk decisions in security notes.
No deletion or archiving is planned for non-fork repositories. Fork-only cleanup is in scope.
| Area | Signal |
|---|---|
| Measured outcomes docs | 6 of 6 selected repos |
| Best checklist coverage | 12/12 |
| Lowest checklist coverage | 12/12 (dragon-portfolio uplifted) |
| External contribution evidence | 1 external PR opened and 1 meaningful issue comment posted (see tracker) |
| Partner-portal baseline closure | CODEOWNERS added; CLAUDE.md present |
| GitHub contributions (last year) | 1,645 |
| Fork cleanup | 4 forked repositories archived |
- Raise
dragon-portfolioproduction checklist score from 4/12 to at least 10/12 with architecture, security, and deployment docs. - Maintain partner-portal baseline (CODEOWNERS and CLAUDE.md now present) and extend release governance evidence.
- Add reproducible benchmark artifacts and trend snapshots for all six pinned repositories.
- Increase release cadence consistency to one semantic tag per flagship repository each month.
- Expand external contribution evidence with merged PR links and concise impact notes every sprint.
- External contribution tracker
- External contribution evidence template
- External PR #1 opened: langchain-ai/langchain#37071
- External issue comment #1 posted: langchain-ai/langchain#31802 (comment)
- External PR #1 issue comment draft
- External PR #1 description draft
- Week 11 measured outcomes
- Week 12 quarterly audit
- Week 12 recap
- 2-phase profile and README plan
- Profile review follow-up (May 2026)
Python, FastAPI, LangGraph, TypeScript, React, Nx, Docker, PostgreSQL, GCP
I am open to collaboration on trustworthy AI tooling, evaluation systems, and production ML platform work.
Preferred collaboration topics:
- RAG faithfulness evaluation and benchmark design.
- MLOps reliability and observability loops.
- Enterprise frontend platform hardening and release governance.
Response expectation: best effort reply within 2 to 5 business days.
- GitHub: https://github.com/Ramdragneel01
- LinkedIn: https://linkedin.com/in/ramprakashdhulipudi