ML researcher building graph-grounded, zero-trust agentic AI for data, cloud, and engineering-document understanding
4
papers here
1
reviewed by Evolve
3
first-author
0
corresponding
—
works · OpenAlex
2
h-index · OpenAlex
Bibliometric figures come from OpenAlex (2026-10-17) and cover the author's entire public record, not just the papers listed here. Reproducibility signal comes from Evolve's own reviews.
Research profile
Output—
No aggregate output score is computed yet: the snapshot provides only raw counts-by-year (1/11/1/4 works for 2023–2026) without a venue-normalized productivity measure. This axis would score total publication volume and its distribution across venues and author roles.
Impact—
Impact cannot be scored yet: the available citation signals (h-index 2, i10-index 0) are too immature for a stable measure. This axis would capture citation accumulation, citing works, and influence on subsequent research.
Focus100/100
A perfect focus score (100) matches the topic distribution: all top topics sit within applied machine learning — topic modeling (7 works), time series analysis and forecasting (5), advanced graph neural networks (4), traffic prediction and management (4), and multimodal machine learning applications (3). The 2026 preprints extend the same graph-centric program into agentic systems and document/diagram understanding, confirming one coherent research line rather than scattered interests.
Collaboration22/100
The collaboration score of 22 reflects an unusually small working circle: only 2 distinct coauthors appear in the recent three-year window, with Venkataramana Runkana and Prathamesh Gadekar as the recurring partners. Three of the four catalog preprints are first-authored, suggesting a compact, first-author-led pattern rather than a broad multi-group network — efficient for speed, but with limited exposure to external teams so far.
Activity83/100
Activity is the strongest axis (83): the snapshot credits 16 works in the recent three-year window, with counts-by-year peaking at 11 works in 2024, dropping to 1 in 2025, and recovering to 4 in 2026 — the year carrying all four arXiv preprints in Evolve's catalog. Sustained first-author output in 2026 indicates the current pace is rising again rather than tapering.
Reproducibility55/100
Reproducibility is scored at 55 from a single reviewed paper (reproPapers: 1), the P&ID reasoning work, rated medium for reproducibility readiness. Evolve's verdict finds the gains convincing — topology QA rising from 36.7–41.3% to 74.3–76.0% — but notes the evaluation is synthetic-only and harness factors are unablated, so the rating rests on one data point and awaits replication on real engineering sheets.
Research summary
Sagar Srinivas Sakhinana is an early-career machine-learning researcher whose indexed work clusters tightly in applied ML: topic modeling (7 works), time series analysis and forecasting (5), advanced graph neural networks (4), traffic prediction and management (4), and multimodal machine learning applications (3). The OpenAlex snapshot credits 16 works in the recent three-year window, with yearly output peaking at 11 works in 2024, and reports an h-index of 2 and an i10-index of 0 — the citation profile of very recent, still-immature output. Evolve's catalog holds four of his 2026 arXiv papers: three first-author preprints on graph, loop, and harness engineering for zero-trust agentic data engineering and analytical processing, agentic cloud engineering, and forward-deployed full-stack engineering for autonomous cloud MLOps; plus a middle-author paper constraining vision-language models with recovered evidence graphs for grounded, faithful P&ID reasoning. That reviewed work reports lifting P&ID topology QA from 36.7–41.3% to 74.3–76.0% on synthetic sheets — a gain Evolve finds convincing, though synthetic-only and with harness factors unablated. Overall, the record describes a focused, prolific builder of graph-grounded and agentic AI systems working within a very small co-author circle (2 distinct coauthors, notably Venkataramana Runkana and Prathamesh Gadekar), whose citation impact has not yet had time to mature.
Collaboration pattern
Collaboration is strikingly compact: the recent three-year window shows only 2 distinct coauthors, with Venkataramana Runkana and Prathamesh Gadekar as the recurring partners, and a collaboration score of 22. Three of the four catalog preprints are first-authored, indicating a small, first-author-led working group rather than a distributed multi-institution network. For external collaborators this means direct access to the core pair and fast iteration, but limited evidence of cross-team or international co-authorship so far.
Research keywords
Highlights
Graph-grounded P&ID reasoning with vision-language models
Zero-trust harnesses for agentic data engineering
First-author 2026 preprint proposing graph, loop, and harness engineering for zero-trust agentic data engineering and analytical processing — the most explicit statement of his graph-and-loop methodology for autonomous data platforms. Not yet reviewed by Evolve.
Agentic cloud engineering with graph and loop patterns
A second first-author 2026 preprint, 'Towards Agentic Cloud Engineering', applies the same zero-trust agent harness and graph/loop engineering ideas to cloud operations, extending the research line toward autonomous infrastructure. Not yet reviewed by Evolve.
Forward-deployed full-stack engineering for autonomous MLOps
A third first-author 2026 preprint on forward-deployed full-stack engineering for autonomous cloud MLOps, rounding out a trio that frames practical deployment engineering itself as a research subject. Not yet reviewed by Evolve.
Timeline
- 2023Earliest indexed output in the OpenAlex snapshot: 1 work, marking the start of the visible publication record.
- 2024Most productive indexed year in the snapshot: 11 works published.
- 2025Indexed output dipped to 1 work — the quietest year in the visible record.
- 2026Resurgence: 4 indexed works, and all four Evolve-catalog arXiv preprints appear — three as first author on agentic data/cloud engineering and autonomous MLOps, plus the P&ID vision-language reasoning paper.
Papers in the Evolve catalog
Frequent co-authors
About this profile
This profile is AI-structured by Evolve.Pub from public bibliographic data (OpenAlex) and Evolve's own paper reviews. It is not written or endorsed by the researcher unless the profile is claimed. Radar dimensions are computed from real bibliometrics with fixed formulas — raw figures are shown alongside; missing dimensions mean insufficient data, never zero. Errors? Claim the profile to correct or hide it.
- Bibliometrics come from an OpenAlex snapshot only; the profile is name-resolved from public bibliographic data, with no ORCID on file to confirm identity.
- Institution and country are missing from the input and left unverified rather than inferred.
- Output and impact dimensions are unscored because citation evidence is insufficient (h-index 2, i10-index 0); read these as early-career signals, not judgments.
- Only one of the four catalog papers carries full Evolve review data (verdict, keywords, reproducibility); the three first-author 2026 preprints are listed without review.