Agile Coaching Portfolio

Venkata Krishna Kalluri (KVK)

What I did, why, and what changed as a result — across four industries and thirteen years.

Agile Coach Enterprise Facilitation Executive Coaching Team Coaching Dependency Mapping Big Room Planning
13 years · Fintech, Banking, Telecom & E-Commerce · Berlin-based

I never told them the answer, I just made it impossible to miss.”

Explore the record
For Leadership

What this means for the organizations I join.

Delivery predictability: dependencies and sequencing surfaced at planning, across 10+ teams, before execution starts.

Escalation load reduced: teams resolve architecture and process problems themselves, in the room.

Decisions compressed: CXO-level metric conflicts and prioritisation disputes closed in one conversation, not chased over weeks.

Proven at scale: a monolith stuck for years got a migration plan from one retrospective — the highest-risk 40% delivered in 3 months.

Core Coaching Capabilities

Every item below is evidenced somewhere in this document.

This portfolio is built the way I coach: evidence first, framework second. Every claim traces to a story, a room, or a number — not an adjective.

ScrumKanbanScaling (Nexus, LeSS, SAFe) Servant LeadershipStakeholder ManagementCXO & Executive Coaching Dependency MappingBig Room PlanningValue Stream Mapping Product CoachingRetrospective FacilitationTeam Coaching & 1:1 Mentoring Outcome-Based MetricsAgile Transformation
01
Business Outcomes

The numbers, at a glance.

An Agile Coach's job shows up in outcomes, not activity. What changed across four companies and thirteen years — organised the way outcomes actually get evaluated: delivery, quality, predictability, and the people side that makes the numbers durable.

Teams Coached
30+
teams across 4 companies
Transformation
2
acquired companies unified, zero shared culture to one product
Release Throughput
+40%
features shipped during Seamless's protected transition window
Defect Rate
-50%
escaped defects halved, same 4-month window
Launch Timelines
-25%
international launch time — redBus, via Big Room Planning
Roadmap Value
+25%
business value delivered — Wells Fargo prioritisation & roadmap coaching
Stakeholder Engagement
+40%
business engagement at Seamless — earned through trust, not mandate
Career Span
13 yrs
5 companies, Scrum Master to Senior Agile Coach

What sits behind each number

Teams coached10+ Scrum Masters mentored, 20–30+ teams coordinated, 200+ professionals trained across Seamless Systems alone; portfolio-wide coaching at Seamless, Wells Fargo; 7 teams at redBus.
Transformation initiativesTwo acquired companies at Seamless Systems, different tech stacks and cultures with zero shared delivery rhythm, brought fully onto iterative ways of working, with team members trained on-site in France and Romania.
Product unification15+ client-specific versions of Seamless's own product consolidated into one, with feature toggles for customisation, over 9 months — reducing OpEx by roughly 30% and removing the need for customer-specific teams.
Delivery improvementsFeatures shipped increased 40% during a protected 4-month transition window at Seamless, once CXO leadership agreed to one metric instead of judging a new way of working against the old one.
Lead time / launch timelinesInternational launch timelines cut 25% across Indonesia, Peru, and Colombia at redBus, through Big Room Planning and continuous reprioritisation on market signal.
Escaped defectsDefects halved in the same 4-month Seamless transition window — the direct result of protecting the team from being measured on old-world velocity while learning a new way of working.
Release frequency & predictabilityShifted teams from output reporting (velocity) to outcome tracking (TTM, NPS, value delivery) at redBus. At Seamless, one agreed metric (iterative delivery) gave CXOs a predictable signal during transition.
Stakeholder satisfactionBusiness engagement up 40% at Seamless. Sprint review attendance recovered after leadership felt the cost of missing decisions. Weekly CEO / daily CTO+CPO cadence at redBus kept trust current, not quarterly.
Dependency visibilityIntroduced structured dependency mapping at three companies — redBus (7 teams), Seamless (4 functions: Product, Engineering, Operations, Sales), and S&P Global — each time cutting cross-team delays measurably.
Individual growth10+ Scrum Masters coached 1:1 into stronger facilitation and stakeholder handling; one Product Manager coached from cold-case remote disengagement, back into the room (see Section 03).
02
Coaching Philosophy & Operating Model

How I actually work, and which stance I'm using.

"I don't hand teams the answer. I design the conditions for them to find it themselves — surface the evidence, ask the question that has no wrong answer, then let the room own the conclusion. That's how resistance turns into ownership: not through argument, but through a diagnosis the team arrives at on its own."

Five things that show up in every engagement

Observe before acting

Sit in Events. Stay quiet. Notice where energy drops, where people go silent. The intervention comes after the diagnosis, not before it.

Find the one willing person

Not the most senior, not the most resistant. Build outward from evidence, one team at a time, until the sceptics run out of reasons.

Make the value visible before asking for commitment

Evidence first. A one-page summary of what changed carries more weight than a pitch for buy-in ever will.

Questions over answers

The team usually knows the problem already; they just haven't said it out loud together. The right question does the rest.

Remove barriers to learning, not the pressure to perform

Change the conditions first — the behaviour follows. Teams can't learn a new way of working while being judged on the old one.

Where the Read Was Wrong

One team delivered roughly 60% of forecast, with items pushed into sprints from unclear sources. New to the team, I read it as a trust problem and spent three to four months on retros and physical-wall visualisation. The actual problem was structural: their working pattern didn't fit Scrum's fixed-sprint model, and the EM had said as much early on: "our team doesn't want to do Scrum." I heard that as resistance, not as data.

Once I recognised the mismatch, we moved the team to Kanban: flow visualised, WIP limits in place, EM and PO owning reprioritisation. Delivery stabilised.

The tools I reached for first were right for a trust problem. This wasn't one.

The stance shifts with what the room needs

"Agile Coach" gets used as one label for four different jobs. I move between them deliberately, and I know which one I'm in at any given moment — that choice is itself part of the craft.

StanceWhat it looks likeWhen I use itEvidence
Coaching Ask the question that has no wrong answer; let the room reach its own diagnosis When a team already has the answer but hasn't said it out loud The jeep retro — architecture decision, team's own conclusion (Section 05)
Mentoring Work alongside one person on their own growth edge, over time Developing Scrum Masters and Product Owners individually, not just as a group 10+ SMs mentored at Seamless; one PM coached into re-engaging stance (Section 03)
Facilitating Design and run the room; stay neutral on content, own the process Multi-team planning, retros, workshops where the group must reach its own output Big Room Planning at redBus — CPO, CTO, CEO, 10 PMs, 10 EMs in one room
Teaching Give direct instruction or a lived simulation when the team lacks a concept, not just a habit New-to-Agile teams, or a concept no amount of questioning will surface on its own The Ball Point Game — self-organisation taught from the inside, not off a slide

A team's behaviour is a symptom of its structure, not its character. I don't fix people; I change what the room can see, and let the system correct itself. That's why the interventions in this document look different from each other (a jeep video, a dependency wall, a protected metric window), but the underlying move is always the same: change the conditions, then get out of the way.

How I operate at the CXO level

Both redBus and Seamless Systems put me in a direct reporting line to the top of the organisation: CEO at redBus, COO (owning both product and technology) at Seamless. That wasn't incidental. An Agile Coach who reports into engineering only sees half the problem. Reporting to outcomes means sitting in the gap between what leadership believes is happening and what the delivery data actually shows, and closing it with evidence, not advocacy.

At redBus

Weekly with the CEO on direction — revisiting product goals against market signal, then translating changes into the roadmap with the CTO and CPO.

Daily Scrum of Scrums with CTO, CPO, and EMs/PMs — run on a physical wall with the full dependency chart visible: what's prioritised, who owns it, ETA, and any blockers.

The gap between the weekly direction conversation and the daily delivery reality was my operating space.

At Seamless Systems

Direct line to the COO — one conversation, both sides of every delivery problem. No translation layer between product and engineering leadership.

Told CXO leadership directly: you cannot run a transformation and a performance review on the same metric. Won a protected 4-month window — and delivered inside it.

03
Individual Coaching

1:1s, Scrum Master growth, and the systemic lens.

The team layer is the visible half of the job. Alongside it, every engagement carries a 1:1 coaching thread — Scrum Masters and Product Owners working through their own blockers, decisions, and growth edges, so the system holds up even when I'm not in the room.

A team maturity lens, not a fixed checklist

Before choosing an intervention, I read where a team actually is — not where the calendar says they should be. I use a simple forming-to-self-managing lens (adapted from Tuckman) as a working diagnosis, not a scorecard I show the team — it shapes which stance I take and which retrospective format I reach for. Two examples of it actually firing, at opposite ends of the scale:

Stage: Forming — Seamless

A team I joined was closed off — a sharp contrast to the openness I'd found at redBus. I read it as a trust gap, not hostility, and started with the lowest-stakes formats I had: Weather Map, Stars & Tweets. No pushing for depth before the team was ready for it. Full arc and outcome in Section 06, Chapter 2 — this is the same team.

Stage: Performing — redBus

A team already strong in iterative delivery didn't need more scaffolding — they needed a stretch. I showed them a Mob Programming video in a retro; I judged them ready for it because they were operating at what I'd call a Performing stage. The team volunteered to run their next sprint story as a mob-programming exercise — their call, prompted by the video, not assigned.

What this looked like at Seamless

Getting Product into the room. Product wasn't in the room, and the PM was remote — a cold case wouldn't work. I found one Product Owner who already had the PM's trust and coached them 1:1 over several weeks: what to say, when to say it, how to bring evidence instead of an argument. I asked that PO to join backlog refinement and sprint Events for one month, no announcement. Friction dropped visibly; that team's feature rollout was faster and cleaner than anything before it. The PO and I went to the PM together to show what had already changed — he agreed to pilot with two more teams. It scaled from there: all POs embedded in engineering delivery rhythms across the whole product organisation within three quarters.

"I didn't convince him. I showed him what had already changed."

This is a 1:1 coaching story before it's a room-level one — the PO's own growth (confidence, evidence-led influence, stakeholder handling) was the actual intervention. The room-level result followed from it.

What this looked like at Wells Fargo

Coaching a peer through her own incentive, not around it. A PM was withholding requirement details from Engineering, and wouldn't say who else on the business side they could go to for clarification — Operations held context the PM wasn't sharing. The root cause wasn't process; it was fear of losing control of the team if she opened that access. I didn't escalate it. I sat with her directly and worked through what withholding access was actually costing her: poor outcomes, and little to show her own manager next quarter. Opening access to Operations meant Engineering would trust her more, not less, and freed her to focus on roadmap and other priorities instead of being the bottleneck for every question. She granted the access. Outcomes improved.

10+ Scrum Masters, mentored individually

04
Career Arc

Scrum Master to Senior Agile Coach.

Five companies, four industries, thirteen years — each stage adding a different layer of scale, regulation, or organisational complexity to the same underlying instinct: make the problem visible, and let the team own the fix.

Maveric Systems
Agile Consultant · 2013–16
Foundation years. Agile learned by doing it with banking clients from scratch, on JIRA and Confluence.
redBus
Scrum Master · 2016–17
Where the philosophy was forged. CEO-direct reporting, CXO coaching, dependency mapping, Big Room Planning, retro experimentation.
Seamless Systems
Agile Coach · 2017–22
Where it met real scale. 10+ Scrum Masters mentored individually, 20–30+ teams. 15+ product versions unified into one; two acquired companies brought onto iterative ways of working.
S&P Global
Agile Coach · 2022
Where it met regulation. Compliance, risk, and delivery in tension. Role ended: org restructure.
Wells Fargo
Agile Coach · 2023–26
Where it met a global bank. Portfolio-wide coaching inside a regulated, matrixed enterprise.
05
Case Study — redBus

The jeep retro & CEO-direct coaching.

RoleScrum Master, direct to CEO
PeriodJuly 2016 – August 2017
SectorE-Commerce, Bengaluru

redBus was the largest online bus ticketing platform in India, in hyper-growth and expanding into Indonesia, Peru, and Colombia — an engineering org of 100+ people where speed mattered as much as reliability.

The centrepiece: the jeep retro

Trigger
B2C platform down 4 hours. Payment glitch. Engineering fixed it. The following week: retro.
What I did
Opened with a 4-minute video — Canadian army soldiers dismantling and reassembling a jeep in field conditions. Asked one question: why do the military use jeeps despite huge budgets?
Why this format
A standard post-mortem after an outage pulls people toward blame and defence. The video put enough distance between the team and the Friday crisis that they could think about the system instead of defending what happened. The question had no wrong answer: the goal was to get them talking about architecture without framing it as architecture.
How the team got there
Product and engineering both in the room. Early answers: cost, simplicity, remote availability. Then: minimal architecture — designed so anyone can fix it, anywhere, fast. Asked them to correlate it to their own system. Kept asking why. Got to the real answer: monolith. Can't isolate. Can't fix fast.
What changed
Action item from the retro: move to microservices. The team's decision, not mine. Payment and booking critical path migrated within 3 months — approximately 40% of the monolith, the highest-risk zone first.
What it didn't fix
The remaining 60% of the monolith stayed as-is. Migrating it wasn't worth the risk against lower-traffic paths, and I said so at the time. Not every insight from a retro should turn into a roadmap item — part of the job is knowing which ones shouldn't.

"I didn't tell them their architecture was wrong. I showed them a jeep. Monolith to microservices. Their call, not mine."

Outcomes at redBus

Architecture

Monolith-to-microservices decision made in retro. Payment and booking critical path migrated in 3 months — team-driven, not top-down.

International Launches

Big Room Planning — CPO, CTO, CEO, 10 PMs, 10 EMs. Rollouts to Indonesia, Peru, Colombia delivered ahead of the 12-month plan through continuous reprioritisation on market signal.

Dependency Mapping

Two-stage: blockers surfaced in planning (nobody could claim ignorance), then lived on a physical wall during delivery — visible daily, impossible to ignore.

Retro Culture

30+ formats: story cubes, darts, draw circles. Teams that used to sit through treat retros out of obligation started asking what was running next.

Leadership Coaching

Weekly with CEO on direction. Daily with CTO and CPO on delivery health — grounded in data, not hope.

06
Case Study — Seamless Systems

One org, three transformations.

RoleAgile Coach, direct to COO
PeriodAugust 2017 – March 2022
SectorTelecom, Hyderabad · ~65 engineers, 7–8 PMs

A different version of the product for every client — separate codebases, separate priorities, separate roadmaps. Product, Engineering, Operations, and Sales never shared a room or a plan. Three chapters, in the order they happened: getting Product into the room, unifying the product line itself, then bringing two acquired companies onto the same way of working.

Chapter 1 — Internal Transformation (Years 1–2)

The first challenge wasn't Agile adoption — it was that Product, Engineering, and Operations ran on separate tracks. Dependency mapping became a first-class practice across all four functions: built collaboratively in planning so nobody could claim ignorance, then lived on a physical wall during delivery, visible and updated daily.

The engineering lead who called Agile a fad

Didn't argue. Found the one person willing to try, ran two sprints with visible output. The lead watched. By the third sprint, the sceptics ran out of reasons — he changed his mind because the team next to his was shipping faster, not because of anything said in a meeting.

"I didn't argue with the sceptic. The results did."

Leadership cancelling sprint reviews

Their mental model was the problem, not their calendar. Sent a one-page summary after every review (what was built, what decision it enabled), then followed up directly. After three cycles, leadership started attending: they had felt the cost of not being there.

"I stopped asking for their time. I showed them what they were missing."

Chapter 2 — One Product, Not Fifteen

One team, when I joined, was closed off — a sharp contrast to the openness I'd found at redBus. I read it as a trust gap and started with the lowest-stakes retro formats I had: Weather Map, Stars & Tweets. Over roughly a quarter, the team opened up (the maturity-lens read behind this choice is in Section 03).

Once that trust existed, I ran Value Stream Mapping with them — a technique that only works if a team hands you the real picture, not a sanitised one. I extended the same exercise with other teams, going deeper with Engineering Managers and Product Managers directly. What it surfaced: the org was carrying 15+ versions of the product, one per client, because every client had asked for different features and nobody had ever stepped back to ask what the actual product was.

I took that finding to my COO, who held both CTO and CPO, and made the business case directly: we were carrying rising OpEx (bug fixes and maintenance duplicated across 15+ versions, different skillsets needed per client) against a CapEx that should have been building one product, not fifteen. That conversation is what authorised the unification effort.

15+
Versions, One Product
9
Months
~30%
OpEx Reduced
5
Eng Teams Involved

With Engineering, Product, Operations, and Sales in the room, I facilitated a product deep-dig using User Story Mapping, User Journeys, and a Product Pruning Tree — mapping where the product carried the most weight, and what could be cut. Engineering data showed which features were actually used and where the real problems sat. Together, that became a roadmap for one generic product. Around 5 engineering teams, 2 Product Managers, and 3 Engineering Managers worked directly on it, while the rest of the org kept serving existing client work that couldn't be paused.

Nine months later: one version of the product, with feature toggles built by Engineering so any client's configuration could be served from the same base — not, as before, a snapshot of whichever client's release happened to be furthest ahead. OpEx dropped by roughly 30%. Customer-specific teams were no longer needed.

"We didn't have a Product. We had fifteen customers, each with their own copy of one."

Chapter 3 — The Acquired Companies (Years 2.5–4.5)

A separate, later challenge — two acquired companies, both running waterfall, different tech stacks, different cultures, and different products from Seamless's own line (these stayed separate products; the goal here was working model, not product convergence). The resistance looked the same in both. Around 50 people each, no reason to trust a new way of working.

CXO metrics pressure

Teams were being judged on old-world velocity while learning a new way of working. Argued directly: you can't run a transformation and a performance review on the same metric. Won one agreed metric (iterative delivery) for a 4-month window. Features rolled out increased 40%, defects halved, time to market improved.

"I argued against the metrics. Got the window. Delivered."

Beyond the room-level resistance, the underlying work was hiring and training: I hired and trained Scrum Masters for both acquired companies, and trained Engineering and Product teams directly on-site in France and Romania. Both companies kept their own products, genuinely different from each other, but fully adopted iterative ways of working by the end of this chapter.

Outcomes at Seamless

Product Unification15+ client versions consolidated into one, with feature toggles for customisation, in 9 months — OpEx reduced roughly 30%, customer-specific teams no longer needed.
CXO MetricsArgued against judging a transformation on old-world velocity. Won a protected 4-month window on one agreed metric — features rolled out up 40%, defects halved.
Acquisitions OnboardedTwo acquired companies, different tech stacks and cultures, brought fully onto iterative ways of working — teams trained on-site in France and Romania.
Leadership EngagementSprint review attendance recovered after leadership felt the cost of missing decisions — one-page summaries after every review, followed up directly.

How I work — what Seamless showed

07
The Facilitation System

30+ formats, a standing practice.

"Make it visible, and it exposes the truth. Make it a game, and a team stops being told what to think — they start working it out themselves."

Thirteen years of coaching taught the same lesson every time: a team fixes what it can see, and it learns fastest when it's playing, not being lectured. This isn't a once-a-sprint ritual; it's a standing system with four mechanisms running underneath every team, every sprint, without exception.

01

Whole-Product Visualization

02

Visible Delivery & Accountability

03

Structured Retrospective Library

04

Purpose-Built Training Formats

How the format is chosen

A sample from the library

StageFormatWhat it doesWhat it produced
Set the StageWeather MapOne placement, sun to storm, before a single word is said.Surfaced mood swings the standup never caught; cut warm-up time by half.
Gather DataTeam Health RadarFive axes — Collaboration, Trust, Accountability, Process, Product — scored live.Surfaced trust issues between Product and Engineering that became the basis for three sprint improvements.
Gather DataRepeat / Avoid / AddA People/Process/Technology matrix forcing every issue to a verdict.Every category walked out with its own action list, not a wall of notes nobody revisits.
Generate InsightsSailboatAnchors for what's dragging the team down, wind for what's pushing it forward.Turned named anchors into actions that helped the team ship faster the next sprint.
DecideValue / Effort QuadrantTeam plots its own backlog by value against effort, agrees live on what's next.Shifted prioritisation from a PO call to a team decision — fewer re-litigated backlogs.
Decide / CloseRetro DartTarget-board prioritisation — closer to centre, higher the rank.Set a clean cutoff for what actually made next sprint's action list.
CloseThank You & My ActionCloses on acknowledgment by name and one personal commitment.Made recognition a habit; every person left with one thing they'd personally committed to.
TrainingThe Ball Point GameA live production simulation — five timed iterations, plan and retro between each.Team voluntarily reduced WIP limits afterward and ran the next sprint with fewer carry-overs.
Also in the toolkit: Sprint View, Stars & Tweets, Super Heroes, 4L's, Fast/Furious/Funny/First/Fed Up, Coin Game, Jenga Blocks, Wall of Experiments & Learnings, Build Your Own Scrum, a Roles & Values card deck, a Value Stream Map exercise, and video-led reframes — the jeep, the IDEO shopping cart, the Nordstrom Innovation Lab, the backwards-brain bicycle — chosen to match wherever the team's head is that week.
08
AI-Augmented Coaching

Practising what I coach.

This isn't a pivot away from coaching. It's a differentiator inside it: every team I coach now is being asked to adopt AI into its own workflow, so I put myself through that exact unfamiliar-change experience first, with no coding background, to know what I'm asking of a team before I ask it.

7
Systems
5
AI Platforms
15
Days
0
Lines Hand-Written

Story Cubes for Scrum

A live, installable PWA for Scrum facilitation. 152 prompts across 14 events, built entirely through AI prompting. Used in the room to coach adoption, not just as a demo.

Life Event Mode (N26 concept)

The most personal build: a contextual banking feature born from my own relocation to Berlin. Full PRD to working prototype, solo, end to end.

Gift & Milestone Hub

A two-agent system grounded in a personal knowledge base (RAG), not generic model output. Parses casual notes into structured profiles, generates localised, budget-matched gift suggestions on request.

My Daily Brief

A manager agent orchestrating four sub-agents across live calendar, transit, and briefing integrations.

AI Build Portfolio

kvk-ai-projects.netlify.app →

Why This Belongs in a Coaching Portfolio

An Agile Coach's credibility with engineering teams comes from having actually built something, not from having read about building. These projects are proof of the same instinct that runs through every case study in this document: don't theorise the change, go through it, then bring back what's true.

Get in Touch

"I never asked anyone to trust the process. I showed them what the process had already done."

Venkata Krishna Kalluri (KVK) — Agile Coach, 13 years across fintech, banking, telecom & e-commerce. Berlin-based.