Building products that create
impact & drive growth.

An internal learning platform at a Tier-1 IT services firm served thousands of consultants across BFSI domains. Phase 1 had role-based learning paths but personalisation didn't exist. Course completion dropped sharply after week two and the L&D team was spending up to 30% of its time chasing follow-ups by hand.

• Discovery was demographic, not skill-based — consultants couldn't find content matched to their actual gaps
• No mechanism for self-directed timelines; learners drifted after the first fortnight
• L&D ops carried the entire follow-up burden manually, burning cycles on chase-ups

• 12 stakeholder interviews across L&D, engineering and BFSI practice leads
• Mapped current vs. target user journey for three personas — junior consultant, mid-level, practice lead
• Designed two flagship Phase 2 capabilities: AI-powered Course Discovery and Milestone Mapping
• Authored full BRD: scope, dependencies, OKRs, acceptance criteria, risk matrix, 14-week 5-phase delivery plan
• Pre-aligned engineering leadership on architecture before formal sign-off — cut formal review cycles in half

• Course Discovery — recommendation engine using role + dept + assessment-based skill gaps
• Milestone Mapping — auto-generated timelines, weekly checkpoints, in-app + email nudges
• Admin reporting layer for L&D to monitor cohort progress at the practice level

At a 5,000+ person services firm, project managers were losing roughly six hours a week to Jira hygiene — drafting acceptance criteria, formatting tickets, generating sprint reports, writing standup notes. High-volume, repetitive, and almost always done in the same way. Classic case for an AI co-pilot built into the existing workflow.

• Junior PMs wrote inconsistent acceptance criteria — engineers asked clarifying questions, cycle time slipped
• Standup notes were duplicated effort: information already lived in ticket transitions
• Sprint retrospectives lost context across sprints — nobody had time to write summaries

• Surveyed 40 PMs across delivery practices to map exact toil patterns
• Prioritised three use cases by frequency × time-saved: AC drafting, standup-from-tickets, sprint summary
• Designed the plugin to surface inside Jira's native UI — no new app, no new workflow, no training required
• Built a lightweight evaluation harness — AC quality scored against an engineer rubric pre- and post-launch

• Draft AC button on every story — LLM generates from the existing description and linked tickets
• Standup digest — auto-generated from ticket transitions in the last 24 hours
• Sprint summary — one-click report covering velocity, blockers, ticket flow

Internal services bench at a 5,000+ person consultancy. Available consultants were matched to incoming client asks via spreadsheets and Slack lookups. Time-to-fit averaged five working days. Mismatches were common — a Java engineer assigned to a Python project, a junior dev to a senior role — and reassignments after week one were a routine cost.

• No structured skill data per consultant — practice managers worked from memory
• No real-time view of who was on the bench and for how long
• Matching depended on tribal knowledge; coverage broke whenever a manager was unavailable

• Defined a controlled skill ontology — 40+ core skills, 5 proficiency levels
• Built a self-service skill profile flow consultants could complete in under 10 minutes
• Designed the matching engine with weighted scoring: skill, level, availability, recency of relevant work
• Shipped MVP with practice managers as the primary user — high-leverage adoption point, not consultant-side

• Consultant skill profile — guided self-entry with proficiency calibration prompts
• Live bandwidth dashboard — practice-level view of who's free, when, with what skill
• Match engine — surfaces top five candidates per request with rationale ("85% match: Python + AWS + recent fintech")