Welcome to my personal corner on the web. I'm Sandeep Tomar. Over the last decade, I've piloted heavy freight locomotives, directed multi-site commercial retail operations, earned an MBA, and coordinated passenger transport networks. Here, I write candid reflections on logistics friction, data-driven decisions, and building with AI.
When you're piloting thousands of tons of rolling stock at speed, momentum is not an abstract physics concept—it is your reality. The mental models forged in the cab of a locomotive apply directly to coordinating regional transit networks and managing frontline operational teams.
By Sandeep Tomar
How non-traditional coders are using modern LLM environments like Cursor to ship bespoke roster checkers, CSV parsers, and operational dashboards in hours.
Why the enterprise frontier has shifted from novelty generative text toward autonomous multi-step agents that execute ground actions and flattened reporting hierarchies.
Exploring how native multimodality processes video CCTV, audio dispatches, and dense technical manuals simultaneously to solve real-world industrial dilemmas.
Analyzing voice-cloned CEO authorizations, automated spear-phishing, prompt injection, and the operational defensive protocols needed to protect facilities.
How agile independent firms and regional operators replicate hundred-person corporate departments in customer support, dynamic pricing, and local inventory forecasting.
Why memorization is dead and critical synthesis is king. Practical frameworks for students using AI as an intellectual sparring partner and prototype builder.
The architectural mechanics of agent loops that observe operational states, generate candidate timetables, execute API calls, and request human-in-the-loop sign-offs.
Implementing zero-retention API endpoints, role-based vector database access, and immutable audit logging to prevent proprietary operational data leaks.
Why retrieval-augmented generation provides instant factual updates for operating manuals, safety circulars, and rosters without expensive retraining cycles.
A four-part financial model calculating direct hours saved, compliance error prevention, and decision velocity against token fees and maintenance TCO.
A three-tier metric framework ensuring ground teams know exactly who is responsible and what countermeasure to take when an operational index slips.
How to turn complex statistical distributions into glanceable heatmaps and actionable procedures that frontline personnel respect and adopt.
Tracking subtle shifts in vibration harmonics, bearing heat, and electrical current draws to replace components before catastrophic failures happen.
Techniques for deduplicating corrupt driver shift records, resolving timezone shifts, and reconciling inconsistent POS till logs into reliable datasets.
Using window functions (LEAD, LAG, PARTITION BY) and CTEs to compute platform dwell times and rolling inventory velocity over millions of rows in seconds.
Why naive threshold alerts cause alarm fatigue, and how exponential moving averages and rolling Z-scores filter true operational emergencies from noise.
Evaluating the architectural cost of event-driven architectures versus scheduled batch aggregations in logistics and multi-site monitoring.
Translating confidence intervals and p-values into Expected Monetary Value (EMV) and risk trade-offs that senior directors readily back.
Why conflicting departmental definitions of 'on-time performance' breed confusion, and how a centralized semantic data layer enforces single-source truth.
Grouping physical assets by commissioning batch to reveal hidden manufacturing defects, component fatigue, and true depreciation curves.
Why momentum on the tracks and operational inertia in business follow the exact same physics: you must preview and brake long before seeing the red signal.
How a 4-minute intermediate dwell delay ripples into 48 minutes of cumulative reactionary delay across intersecting passenger transit corridors.
Replacing physical trackside lights with in-cab moving block signalling to unlock 40% more train paths without laying a single extra mile of physical rail.
The nocturnal rhythm of cleaning, pantograph checks, and wheel turning between 11 PM and 5 AM to guarantee 100% morning passenger service turnout.
Why commuters evaluate door-to-door journey times rather than platform speeds, and how synchronized ticketing creates frictionless public transit.
Decisive triage when overhead lines fail: isolating electrical sectors, rerouting mainline services, and deploying bus replacement contingencies calmly.
Pragmatic solutions for regional lines: discontinuous overhead wires, battery-electric multiple units, and regenerative braking power recapture.
Formulating timetable compilation as a combinatorial optimization problem balancing platform availability, driver breaks, and resilience buffers.
How intelligent platform zoning, clear passenger door guides, and step-free flow eliminate boarding bottlenecks during commuter peak hours.
Managing headway separations, loop passing tracks, and speed differentials between 2,000-ton freight convoys and 100 mph commuter services.
Balancing daily fuel audits, fast customer velocity, and strict compliance when footfall never stops. Practical leadership lessons from the frontlines.
How principles of cyclic fatigue stress, thermal dissipation, and mechanical dampening translate into team burnout buffers and corporate balance sheets.
Why top aviation and rail organizations treat near-misses as urgent alarms and ensure ground competence overrides titles in operational crises.
Designing forward-rotating circadian schedules that safeguard driver alertness and staff morale while guaranteeing 24/7 coverage.
Why hyper-lean Just-in-Time models break under pressure, and how strategic buffer stocks and operational slack prevent cascade failures.
Establishing transparent KPI audit scorecards and regular review cadences to transform contentious supplier dealings into reliable alliances.
Mastering the 'Three Whats' of emergency leadership: What do we know? What are we doing right now? What do we need from you?
Why organizational excellence is built through hundreds of small frontline improvements rather than top-down sweeping overhauls.
Rigorous cycle counting, goods-received inspections, and shrinkage root-cause investigations to protect working capital in retail.
Active listening, visible leadership on the shop floor, prompt resolution of grievances, and recognizing reliable operational discipline.
I've always been drawn to how physical systems move. Long before I worked behind dashboards or strategic spreadsheets, I sat inside the cab of heavy locomotives on vast rail networks. In that seat, safety isn't an abstract corporate presentation; it is every single signal aspect, every brake test, and every fraction of a second when judging deceleration distances in heavy weather.
That foundational engineering discipline—grounded in my Mechanical Engineering degree—instilled a lifelong habit: respect the physics of operations. Systems fail not when people make honest mistakes, but when workflows lack built-in friction buffers.
Later, moving into multi-site commercial retail and high-volume forecourts in the UK, I encountered a very different kind of operational flow: human footfall, supply chain deliveries, inventory rotas, and daily financial audits. Here, speed and morale are everything. It prompted me to pursue my MBA at an accredited British university, marrying operational instincts with financial acumen and organizational strategy.
Today, coordinating passenger rail and regional transport services while integrating Google Data Analytics and modern AI tooling (Cursor AI, Python, spreadsheets), I spend my spare time synthesizing these experiences into essays. My goal with this blog is simple: share grounded, practical perspectives for anyone managing complex systems, frontline teams, or logistics challenges.
Never sacrifice ground truth or safety protocols for administrative speed.
Data without operational empathy is just clutter. Always interrogate the source.
Adopt AI and code where it reduces friction, not where it adds unnecessary layers.
A snapshot of the software, analytical tools, and resources I use daily to analyze data, streamline operations, and write.
For building lightweight operational scripts, automated parsing routines, and interactive dashboards rapidly.
The timeless backbone of operations analytics. Querying delay logs, inventory turns, and revenue variances.
Real-time railway network monitoring, delay attribution software, and incident response matrices.
Transforming complex analytical insights and safety bulletins into memorable, glanceable visual guides.
I welcome constructive feedback, shared operational dilemmas, and thought-provoking discussions.