Mayank Roy · candidate brief · Production Solutions Engineer, North America

Petroleum engineer who builds the automation

Seven-plus years on Permian wellsites, an MS in Petroleum Engineering, and the Python, SQL, Power BI and AI-agent skills to turn an operator's own data into a Proof-of-Value they can act on in weeks, not quarters.

Permian wellsite experience+Python · SQL · Power BI+AI agents→faster Proof-of-Value for operators
See the live demo → Resume (PDF) Email
MS Petroleum Engineering, New Mexico TechDelaware & Midland Basins, 2017–2025U.S. Permanent Resident, no sponsorship neededSan Antonio, TX · open to relocation

The job, line by line

Each duty in the posting, and what I have already done that matches it

DutyMy evidenceProof
01Understand customer production challenges
Seven-plus years on horizontal Bone Spring and Wolfcamp wells in Eddy and Lea counties, NM and the Midland Basin. Senior man on the unit, the technical contact between rig floor, operator geologist and company man through 12-hour tours.EDGE Services, Weatherford, Paladin, Selman & Associates · MS Petroleum Engineering
GapFinder: how I frame lost production
02Deliver demos and Proof-of-Value
Built GapFinder on public NM OCD data: Arps fits for 36 Delaware Basin wells, the gap split into availability, operating, design and reservoir causes, priced and ranked, with an AI-written morning report. Also run two live demos of operations and finance automation.Independent work, 2025–present
gapfinder.agentroy.xyzoil.agentroy.xyz
03Shape solution strategy with sales
Took a client from problem to production: a CPA firm's raw bank and card exports now become coded, import-ready ledger entries. Rules handle the routine, AI handles only the new items, and each decision is written back as a rule, so review volume falls every month.Client engagement, Python + SQL Server
Sample PoV proposalcpa.agentroy.xyz
04Build customer relationships
Built and maintained Power BI dashboards over real-time drilling and geological data for operator clients, engineering teams and management. As QA manager, coordinated field crews, office teams and clients and drove corrective actions to closure.EDGE Services 2023–2025 · GeoTek USA 2025
Resume
05Run workshops, share thought leadership
Trained mud loggers and technical staff on every system I delivered, and documented the data standards behind them. Wrote a field-grounded piece on where AI agents help production engineers and where humans must stay in the loop.EDGE Services · article on pov.agentroy.xyz
Read the article
06Feed the field back to product
Turned field pain into tools: CompanyMan, in development, reads photographed tank boards and mud reports and returns a formatted daily report. On the unit I caught depth, lag and sensor discrepancies at the source and standardized how crews captured data so it could be combined.cm.agentroy.xyz · EDGE Services
cm.agentroy.xyz

Field to software

Software first, then seven-plus years in the field, now both at once

Plain version: I have written production SQL, sat on the rig floor at 3 a.m. watching a gas show, and checked contractors' work against the plan. That is the mix a production solutions role draws on.

  1. 2012 – 2014
    Senior systems engineer
    Infosys, India
    SQL and PL/SQL against production systems, under change control and code review.
  2. Graduate school
    MS Petroleum Engineering
    New Mexico Tech, Socorro
    Engineering fundamentals behind the field years that followed.
  3. 2017 – 2025
    Permian wellsite operations
    Selman, Weatherford, Paladin, EDGE Services
    Wellsite geologist and mud logger, then senior unit supervisor and Power BI developer on Delaware and Midland Basin horizontals.
  4. Aug – Nov 2025
    Quality Assurance Manager
    GeoTek USA
    On-site inspections against plan and standard, audit-ready records, corrective actions to closure.
  5. 2025 – now
    AI automation engineer
    Independent (Agentroy)
    Python and SQL pipelines, AI agents that handle the routine and escalate exceptions, deployed on my own Linux servers.

Swipe sideways to see the whole line.

How I would run a Proof-of-Value

Thirty days, four steps, one number the operator can check

DAYS 1–7Baseline data DAYS 8–14Gap classification DAYS 15–21Recommended actions DAYS 22–30Value case
Step 1Baseline data

Monthly and daily production, well tests, downtime logs and lift settings for the agreed wells. Fit a decline per well so "expected" is the well's own history, not a guess.

Step 2Gap classification

Split expected minus actual into availability, operating, design and reservoir causes, with rules the operator's engineers can read and argue with.

Step 3Recommended actions

A ranked list with an owner, a cost and a first check for each well. Cheap, recoverable fixes first; reservoir changes go to the forecast, not to a workover.

Step 4Value case

Barrels per day recovered on the wells actually worked, times net price, in $/month. Plus engineer hours saved per week. Both checked against the operator's own allocation.

Done once already on public data: across 36 Delaware Basin wells, GapFinder found 425 bbl/d of gap and about 200 bbl/d that is realistically recoverable, worth roughly $464K a month at $75/bbl. See how it was built.

Tools I use

Pythonpandas, SciPySQLSQL Server, PostgreSQL, MySQLPower BIDAX, Power Query Decline-curve analysisArpsLLM agentsWorkflow automationData QA at the source Excel + VBALinux, nginx, GitPason / WITS rig data

Contact

Happy to walk through GapFinder on a call, or run it on your sample wells

Fastest way to reach me is email.

Emailroy.mayaank@gmail.com
LinkedInmayank-roy-9329721b0
ResumeMayank_Roy_Resume.pdf
Demosgapfinder · pov · agentroy.xyz