Amey-Thakur
- 1k skills
- 0 followers
- 2 days ago last updated
- ▌ Utility CSS · amey-thakurUse Tailwind-style utilities with discipline, extracting components at the right threshold and keeping class lists readable. Use when working in a utility-first codebase or deciding whether and how to adopt one.
- ▌ Orm Tradeoffs · amey-thakurUse ORMs for their productivity while avoiding N+1 queries and knowing when to drop to raw SQL. Use when working with an ORM or debugging the performance problems ORMs quietly cause.
- ▌ Schema Design · amey-thakurDesign database schemas that stay correct under growth and change. Use when creating tables, modeling relationships, or planning schema migrations.
- ▌ Observability · amey-thakurInstrument software so production questions get answered from signals, not guesses. Use when adding logging, metrics, tracing, or alerts, or when a system is hard to debug in production.
- ▌ Profiling Cpu · amey-thakurFind the code that actually burns CPU time using a sampling profiler and a flame graph instead of guesswork. Use when a program is slower than it should be and you need to locate the hot path before touching any code.
- ▌ Containerization · amey-thakurWrite container images and compose setups that are small, reproducible, and safe to run. Use when writing Dockerfiles, docker-compose files, or debugging container behavior.
- ▌ Design Systems · amey-thakurBuild and maintain a design system that keeps a product visually consistent and fast to build. Use when creating design tokens, component libraries, or UI guidelines.
- ▌ Agent Workload Balancing · amey-thakurRedistribute work across desks and workers as load shifts, so one bottleneck does not stall the company while others idle. Use when some desks are overwhelmed and others have nothing queued.
- ▌ Virtual Engineering Team · amey-thakurRun a product-engineering squad of agents from spec to shipped change with a PM, an architect, two engineers, a reviewer, and QA. Use when a task is too large for one agent and needs both division of labor and an independent quality gate.
- ▌ Golden Path Templates · amey-thakurProvide starting templates that encode the organisation's defaults so new services are consistent without anyone reading a standards document. Use when new services vary widely in structure and quality.
- ▌ Platform Abstractions · amey-thakurChoose the level of abstraction a platform exposes so it hides complexity without hiding what teams need to control. Use when designing platform interfaces or when teams keep bypassing them.
- ▌ Cross Team Coordination · amey-thakurKeep work moving across team boundaries with explicit interfaces, owners, and shared checkpoints. Use when a project spans teams and progress stalls at handoffs.
- ▌ Project Risk Management · amey-thakurIdentify what could derail a project, decide responses in advance, and track them rather than reacting when they happen. Use at project start and at every major checkpoint.
- ▌ Automation Guardrails · amey-thakurPut confirmation gates, blast-radius limits, audit trails, and kill switches around automation that can destroy things. Use when building scripts or bots with destructive power.
- ▌ Powershell Essentials · amey-thakurWrite PowerShell that leans on the object pipeline, handles errors deliberately, and runs cross-platform where needed. Use when automating Windows or writing pwsh scripts for mixed fleets.
- ▌ Open Source Review Board · amey-thakurRun an open source review board that gates incoming dependencies on license compatibility, sets the policy for contributing back, and requires a security review before adoption. Use when an organization needs consistent control over which OSS it pulls in and what it publishes.
- ▌ Business Model Design · amey-thakurDecide how the business creates, delivers, and captures value, and check the pieces are consistent with each other. Use when starting something, or when growth is not converting into a viable business.
- ▌ Operations Management · amey-thakurRun repeatable delivery reliably by finding the constraint, reducing variability, and measuring flow rather than utilisation. Use when delivery is inconsistent or slower than the work itself requires.
- ▌ Organizational Design · amey-thakurStructure teams, ownership, and decision rights so work flows rather than queueing at handoffs. Use when coordination cost is rising faster than headcount.
- ▌ Algorithmic Complexity · amey-thakurReason about how running time and memory grow with input size, and know when the asymptotic answer is the wrong one. Use when choosing an approach, reviewing code that loops over data, or explaining why something slows at scale.
- ▌ Concurrency Primitives · amey-thakurUse locks, atomics, channels, and barriers correctly, and know which class of bug each prevents. Use when writing code that shares state between threads.
- ▌ Docs Information Architecture · amey-thakurOrganize documentation with the Diataxis four-quadrant model, navigable structure, and searchability. Use when structuring a docs site or when readers cannot find what they need.
- ▌ Onnx Export Pipelines · amey-thakurExport a trained model to ONNX that runs portably across runtimes, closing operator gaps and proving numeric parity against the source framework. Use when a PyTorch or TensorFlow model must run outside its training stack and the export must be trusted, not just produced.
- ▌ Tensorrt Optimization · amey-thakurCompile a trained model into a fast, GPU-specific TensorRT engine by controlling precision, defining dynamic shape profiles, and proving the fused engine kept its accuracy. Use when a PyTorch or ONNX model must reach a hardware latency floor that eager execution cannot.
- ▌ Typescript Strictness · amey-thakurTurn on and migrate to TypeScript's strict flags incrementally so the compiler catches real bugs. Use when configuring TypeScript strictness or tightening a loose codebase without a big-bang rewrite.
- ▌ Curriculum Sequencing · amey-thakurOrder material so each piece is learnable when it arrives, with prerequisites satisfied and difficulty increasing gradually. Use when designing a course, onboarding path, or tutorial series.
- ▌ Feedback For Learning · amey-thakurGive feedback that improves performance, timed and framed so it is used rather than defended against. Use when coaching, reviewing work, or mentoring.
- ▌ Learning From Failure · amey-thakurExtract the lesson from a mistake by examining the reasoning that produced it, rather than only the outcome. Use after something goes wrong, personally or as a team.
- ▌ Mental Model Building · amey-thakurBuild an internal model of how a system works so you can predict its behaviour rather than recalling procedures. Use when learning a complex system or when you can follow steps but cannot debug.
- ▌ Agent Accountability Loop · amey-thakurTrack every commitment from assignment to completion, chasing what is late and escalating what is stuck, so work finishes rather than fading. Use when things are agreed in meetings and quietly never happen.
- ▌ Agent Delegation Protocol · amey-thakurDecide what to hand to an agent, what to keep with a human, and what to split, based on reversibility and consequence rather than capability alone. Use when deciding how much of a workflow to automate.
- ▌ Agent Design Review Panel · amey-thakurConvene several reviewer agents with fixed distinct lenses, correctness, security, cost, and simplicity, then synthesize their verdicts into one. Use when a single review agent keeps missing whole classes of problems because it cannot hold every concern at once on a design or plan.
- ▌ Agent Generate And Verify · amey-thakurPair every generating agent with an independent verifier whose only job is to refute the output, so plausible-but-wrong work is caught before it ships. Use when an agent's output will be acted on and being wrong is expensive.
- ▌ Agent Plan Execute Replan · amey-thakurSeparate planning from execution and replan when reality diverges, so a long task does not follow a plan that stopped being true. Use when work has many steps and early assumptions are likely to be wrong.
- ▌ Agent Security Audit Team · amey-thakurRun a security audit as agents that map the attack surface, review each surface in parallel, skeptic-test every finding, and assemble one report. Use when you want a codebase audited by a coordinated agent team instead of a single review pass.
- ▌ Revenue Recognition Basics · amey-thakurUnderstand why cash collected is not revenue earned, and record deferred revenue so financial reporting reflects delivery. Use when building billing data models or reporting on a subscription business.
- ▌ Project Status Reporting · amey-thakurReport progress in a form that surfaces problems early and lets a reader act, rather than reassuring. Use when reporting to sponsors or across teams.
- ▌ Realtime Permissions · amey-thakurEnforce and revoke access on live connections so a permission change takes effect immediately rather than at the next page load. Use when shared documents have varying access levels.
- ▌ Search Result Presentation · amey-thakurPresent results so a user can judge relevance without opening each one, with useful snippets, clear grouping, and honest metadata. Use when results are correct but users still cannot find what they need.
- ▌ Architecture Review Board · amey-thakurRun an Architecture Review Board that filters proposals by blast radius, reviews them through fixed lenses, issues binding decisions recorded as ADRs, and honors an appeals path. Use when a technical choice crosses team boundaries and needs a durable, accountable decision instead of hallway consensus.
- ▌ Accessibility Specialist Role · amey-thakurOperate as an accessibility specialist who audits against WCAG, trains teams to stop shipping the same defects, and owns the compliance sign-off. Use when a product needs to be usable with a keyboard, a screen reader, or magnification and someone must certify it before launch.
- ▌ Stakeholder Management · amey-thakurIdentify who is affected by a decision, what each needs, and how to keep them informed enough to stay aligned. Use when a project crosses teams or when decisions keep getting reopened.
- ▌ Floating Point Behavior · amey-thakurReason about rounding, precision limits, and comparison in floating point so numerical results are predictable. Use when money, physics, aggregation, or any comparison of computed values is involved.
- ▌ Operating System Basics · amey-thakurUnderstand what the kernel does for your process, so system-level behaviour such as scheduling, file descriptors, and signals stops being mysterious. Use when debugging behaviour that application-level reasoning cannot explain.
- ▌ Gpu Cluster Scheduling · amey-thakurSchedule GPU jobs on a shared cluster so distributed jobs get all their GPUs at once, fragmentation stays low, and preemption is predictable. Use when several teams share a GPU pool and jobs sit pending while GPUs sit idle.
- ▌ Translation Quality Review · amey-thakurReview translated text for accuracy, register, and fit in context, with a rubric rather than an impression. Use when accepting translations or diagnosing why a localised product feels wrong to native speakers.
- ▌ Self Directed Learning · amey-thakurLearn something substantial without a course, by defining the goal, finding the right resources, and building feedback into the process. Use when you need a capability nobody is going to teach you.
- ▌ Agent Blackboard Workspace · amey-thakurCoordinate agents through a shared written workspace they each read and update, instead of passing messages. Use when several agents contribute to one artefact over time and message passing loses state.
- ▌ Agent Iterative Refinement · amey-thakurImprove output through repeated critique and revision against a fixed standard, stopping when the standard is met or gains stop. Use when first-pass output is close but not good enough and the quality bar is writable.
- ▌ Agent Loop Until Exhausted · amey-thakurKeep running discovery rounds until several consecutive rounds find nothing new, rather than stopping at a fixed count. Use when searching for an unknown number of items such as bugs, risks, or edge cases.
- ▌ Agent Tournament Selection · amey-thakurGenerate several independent attempts, score them against explicit criteria, and select or combine the winner, instead of iterating one attempt. Use when the solution space is wide and the first approach may not be the best.
- ▌ Email Deliverability · amey-thakurGet email into the inbox rather than the spam folder, through authentication, reputation, and list hygiene. Use when email is not arriving or engagement suddenly drops.
- ▌ Notification Fatigue · amey-thakurDetect and reverse the point where users stop reading, by measuring engagement decay and cutting volume before they mute. Use when open rates are falling or mute rates are rising.
- ▌ Local Development Setup · amey-thakurMake a working local environment reachable in minutes from a clean machine, and keep it close to production. Use when onboarding takes days or developers cannot run the system locally.
- ▌ Prioritization Frameworks · amey-thakurPrioritize with RICE-class scoring applied honestly, cost of delay, and refusals that carry reasons. Use when ordering a roadmap or defending priority calls under stakeholder pressure.
- ▌ Optimistic UI Updates · amey-thakurApply changes locally before the server confirms, then reconcile or roll back honestly when it does not. Use when latency makes a correct-but-slow interface feel broken.
- ▌ Cost Structure Analysis · amey-thakurUnderstand which costs are fixed, variable, and step, so you know how profit responds to volume and where cuts actually help. Use when planning capacity, cutting costs, or explaining why margin is not improving with growth.
- ▌ Compilers And Toolchains · amey-thakurUnderstand what happens between source and running program, so build errors, linking failures, and optimisation surprises become tractable. Use when builds fail obscurely or behaviour differs between debug and release.
- ▌ Data Structure Selection · amey-thakurChoose the structure whose operations match how the data will actually be used, rather than defaulting to a list or a map. Use when designing anything that stores and queries data in memory.
- ▌ Fault Tolerant Training · amey-thakurKeep a long training job alive across GPU failures, node evictions, and stragglers so one bad host costs minutes, not the whole run. Use when a run spans enough GPUs and hours that hardware failure during the job is expected, not hypothetical.
- ▌ Kernel Profiling Nsight · amey-thakurProfile GPU work with Nsight Systems and Nsight Compute to read the timeline, name the bottleneck class, and pull the metric that dictates the fix. Use when a GPU program is slower than expected and you need evidence before touching a kernel.
- ▌ Quantization Deployment · amey-thakurQuantize a trained model to INT8 or INT4 for inference, calibrate the ranges, and gate the release on a measured quality regression. Use when serving needs lower latency and memory and you will spend effort keeping accuracy inside a defined budget.
- ▌ Tensor Core Utilization · amey-thakurGet matrix multiplies onto the tensor cores by fixing shapes, precision, and alignment, then measure that the cores actually fired. Use when a GEMM or attention kernel runs far below the card's advertised throughput and you suspect it fell back to the CUDA cores.
- ▌ Triton Inference Server · amey-thakurDeploy models on NVIDIA Triton with a valid model repository, ensemble pipelines, and concurrent execution tuned to keep the GPU saturated. Use when serving one or more models through Triton and configuring layout, dynamic batching, instance groups, or a preprocess-plus-inference pipeline.
- ▌ Explaining To Beginners · amey-thakurExplain something to someone with no background, without condescension and without jargon that quietly requires prior knowledge. Use when writing introductory material or onboarding someone new.
- ▌ Agent Customer Success Team · amey-thakurRun retention as a standing function with agents that watch usage signals, spot churn risk early, prepare account reviews, and hand humans a specific reason to reach out. Use when customers leave without warning and renewals are handled reactively.
- ▌ Notification Batching · amey-thakurGroup related events so a burst of activity produces one message instead of twenty, with rules that stay predictable. Use when a single action generates many notifications or activity comes in bursts.
- ▌ Notification Strategy · amey-thakurDecide what deserves a notification, on which channel, and at what urgency, before building any of it. Use when notifications are added per feature and users are starting to mute everything.
- ▌ Environment Provisioning · amey-thakurCreate development, preview, and staging environments that resemble production closely enough to be useful, and clean them up automatically. Use when environments drift, cost too much, or block testing.
- ▌ Schedule And Critical Path · amey-thakurSequence work by dependency to find which tasks actually determine the finish date, and manage those. Use when a project has many parallel workstreams and an immovable date.
- ▌ Conflict Resolution UX · amey-thakurPresent merge conflicts to users so they can resolve them confidently, rather than resolving silently and losing work. Use when concurrent edits cannot be merged automatically.
- ▌ Presence And Awareness · amey-thakurShow who is here, where they are, and what they are doing, at a cost that scales and without leaking more than intended. Use when adding cursors, avatars, or typing indicators to a shared surface.
- ▌ Production Readiness Review · amey-thakurRun a production readiness review that checks SLOs, runbooks, capacity, and dependencies before an on-call team agrees to own a service. Use when a service is about to be onboarded to a support rotation or handed from its builders to operators.
- ▌ Negotiation Fundamentals · amey-thakurPrepare and conduct a negotiation around interests and alternatives rather than positions, so agreements are better and hold. Use before any commercial conversation with something at stake.
- ▌ AI Datacenter Networking · amey-thakurLay out the network for distributed training so collectives run on the fastest link that spans them, using NVLink, InfiniBand, and topology-aware placement. Use when a training job spans multiple GPUs or nodes and interconnect, not compute, is capping throughput.
- ▌ Agent Incident Response Team · amey-thakurReplicate a live incident team as agents with a commander, parallel investigators, comms, and a scribe, coordinated on a fixed cadence. Use when you want an outage worked by a coordinated agent team instead of one agent debugging alone.
- ▌ Agent Progressive Disclosure · amey-thakurGive agents context in stages, escalating to more capable models or fuller context only when the cheap path proves insufficient. Use when most tasks are easy but a few are hard and cost or latency matters.
- ▌ Quiet Hours And Timing · amey-thakurSend at times that respect the recipient's local hours and working pattern, deferring anything non-urgent. Use when notifications reach people at night or on weekends.
- ▌ Binary Data Representation · amey-thakurWork with bytes, endianness, alignment, and binary formats so data written by one system is read correctly by another. Use when parsing binary protocols, file formats, or debugging corrupted data.
- ▌ Transactional Messaging · amey-thakurSend the messages a user is waiting for, such as receipts, resets, and confirmations, with reliability and clarity that marketing messages do not need. Use when building password resets, receipts, or order confirmations.
- ▌ Comments And Annotations · amey-thakurAttach comments to content that keeps changing, so anchors survive edits and threads stay resolvable. Use when adding review, feedback, or discussion to a document or design.
- ▌ Gpu Utilization Monitoring · amey-thakurMonitor a GPU fleet so you can tell a genuinely busy GPU from one reporting 100 percent utilization while computing almost nothing, and find the wasted spend. Use when GPUs look busy on the dashboard but throughput or cost per token says otherwise.
- ▌ Mixed Precision Deployment · amey-thakurShip FP16, BF16, or FP8 training and inference that holds accuracy while capturing the speedup, using loss scaling and numeric validation. Use when moving a model off FP32 to run faster or fit in less memory and the result must stay correct.
- ▌ Internal Developer Platform · amey-thakurBuild a platform that lets product teams ship without filing tickets, treating the platform as a product with users. Use when infrastructure work is a bottleneck or every team solves the same problems differently.
- ▌ Self Service Infrastructure · amey-thakurLet teams provision what they need within guardrails, without waiting for an infrastructure team to act. Use when provisioning requests queue and infrastructure work is reactive.
- ▌ Security Development Lifecycle · amey-thakurApply Microsoft's Security Development Lifecycle so threat modeling, tooling, and sign-off are built into each phase instead of bolted on before ship. Use when building or shipping software that takes untrusted input or handles sensitive data and you need a repeatable security process, not a one-time audit.
- ▌ Difficult Customer Conversations · amey-thakurHandle angry, distressed, or unreasonable customers without escalating the conflict or capitulating on things you cannot give. Use when a conversation has become emotional rather than transactional.
- ▌ Teaching Technical Concepts · amey-thakurExplain a technical idea by starting from what the learner knows, one concept at a time, with a concrete example before the abstraction. Use when teaching, mentoring, or writing an explanation.
- ▌ Agent Competitive Analysis Team · amey-thakurRun competitive analysis as agents that tear down assigned competitors on one template, enforce sourcing rules, and synthesize implications for your product. Use when you want a market landscape built by a coordinated agent team rather than one agent skimming marketing sites.
- ▌ Live Cursors And Selection · amey-thakurRender other people's cursors and selections so they inform rather than distract, and stay correct as content changes. Use when showing where collaborators are working in a shared document.
- ▌ Checkpointing Large Training · amey-thakurCheckpoint multi-node training runs so a save costs seconds instead of minutes and a resume reproduces the run exactly. Use when a job is large enough that a crash without a recent, verified checkpoint means losing hours of GPU time.
- ▌ Distributed Training Scaling · amey-thakurScale training from one GPU to many by moving through data parallel and sharded FSDP modes, overlapping communication with compute, and reading the efficiency curve to find the ceiling. Use when adding GPUs stops making training proportionally faster and you need to locate where the scaling leaks.
- ▌ Agent Orchestration Antipatterns · amey-thakurGuard a multi-agent team against the structural failures single agents never hit: context lost across handoffs, false agreement, and runaway cost. Use when a team costs more than one agent or trusts its own output, and you need to find out why.
- ▌ Financial Statements Literacy · amey-thakurRead a profit and loss, balance sheet, and cash flow statement well enough to see what a business is actually doing. Use when reviewing your own numbers, a partner's, or an acquisition target's.
- ▌ Multi Channel Orchestration · amey-thakurCoordinate across push, email, in-app, and SMS so a user gets one message rather than the same thing four times. Use when several channels exist and users receive duplicates.
- ▌ Unsubscribe And Preferences · amey-thakurGive users real control over what they receive, with granular options and an honest one-click exit. Use when building preference centres or reducing complaint rates.
- ▌ Developer Productivity Metrics · amey-thakurMeasure the system's throughput and stability rather than individual output, and use the numbers to remove friction. Use when improving engineering effectiveness or justifying platform work.
- ▌ Collaborative Editing Models · amey-thakurChoose between operational transformation, CRDTs, and locking for multi-user editing, and understand what each costs. Use when more than one person can edit the same document at the same time.
- ▌ Inference Serving Optimization · amey-thakurTune LLM serving to hold latency SLOs while raising GPU throughput, working the batch scheduler, KV cache, and paged attention together. Use when a serving replica misses its latency target or leaves memory and utilization on the table.
- ▌ Push Notification Permissions · amey-thakurAsk for push permission at a moment when the value is obvious, and handle denial gracefully, because the prompt only comes once. Use when adding push to a mobile or web product.