• Format — 1,500+ free practice MCQs across 12 top-level topics, with instant explanations after every wrong answer.
• Duration — about 17 hours of focused practice; most learners spread it over 3–5 weeks.
• Level — advanced. Assumes solid recursion, BFS/DFS, and data-structure fluency from Parts 1 and 2.
• Cost — completely free, including the verifiable completion certificate.
• Audience — FAANG and product-company interview candidates, GATE CSE aspirants, competitive programmers chasing higher ratings, working developers preparing for senior or staff-level interviews.
• Prerequisites — finish Parts 1 and 2 of the DSA Mastery series, or have equivalent comfort with recursion, BFS/DFS, and the underlying data structures.

Four audiences. First, FAANG and product-company interview candidates targeting senior roles where DP and graph questions appear in 30–50 percent of onsite rounds. Second, GATE CSE aspirants — the Algorithms section is dominated by DP recurrences and graph algorithms (shortest paths, MST, topological sort, SCC), and this course covers all of them. Third, competitive programmers stuck in the 1400–1800 Codeforces band where the breakthrough is fluency on DP state definitions and advanced graph techniques. Fourth, working developers preparing for senior or staff-level interviews where deep algorithmic problem-solving is the differentiator. The course assumes Parts 1 and 2 of the DSA Mastery series — if you're not solid on recursion and BFS/DFS, go back to Part 2 first.

The curriculum is organised into three layers that map directly to the 12 topics. You can take them in order or jump to whichever layer matches your current gap.

DP foundations

• Introduction to DP & Graph Algorithms — the signals that tell you a problem is DP (overlapping subproblems, optimal substructure, manageable state space), the signals that tell you it's a graph problem (entities with relationships, transitions, reachability), and the boundary between them.
• DP Fundamentals — state definition (the hardest step in any DP problem), recurrence construction (writing the transition between states), and the memoisation-vs-tabulation trade-off including when each is preferred.
• 1D Dynamic Programming — linear DP patterns, binary-decision DP (take-or-not-take), and the classic 1D problems (Fibonacci, climb stairs, house robber, coin change, maximum subarray).

DP patterns

• 2D Dynamic Programming — grid DP (paths, obstacles, minimum path sum), two-index DP for pairs of sequences, and the space-optimisation patterns that drop O(n²) memory to O(n).
• Knapsack Family — 0/1 Knapsack, Unbounded Knapsack, and the canonical variants: partition equal subset sum, target sum, rod cutting, coin change ways.
• Sequence DP — the LIS family (Longest Increasing Subsequence and variants), the LCS family (Longest Common Subsequence and variants), and the Edit Distance family (Levenshtein, edit-distance variants with custom costs).
• Interval, Tree & Bitmask DP — Interval DP (matrix chain multiplication, palindrome partitioning), Tree DP (root-to-leaf and subtree-aggregate problems), and Bitmask DP for state spaces small enough to enumerate by bitmask (Travelling Salesman intro, subset-state problems).

Graph algorithms

• Shortest Paths — single-source shortest path with Dijkstra (non-negative weights), Bellman-Ford (handles negative edges, detects negative cycles), and all-pairs shortest path with Floyd-Warshall. Plus shortest-path variants — k-shortest paths, shortest path with constraints.
• Minimum Spanning Tree — MST concepts (cut property, cycle property), Prim's algorithm (priority-queue-driven), and Kruskal's algorithm (sort-and-union-find-driven).
• Topological Sort & Union-Find — topological sort for DAG ordering and dependency resolution, DAG algorithms (longest path, counting paths), and the Union-Find data structure (path compression, union by rank) that powers Kruskal's and connectivity queries.
• Advanced Graph Algorithms — Strongly Connected Components (Kosaraju, Tarjan), bipartite-graph checks and graph colouring, and articulation points and bridges for connectivity analysis.
• DP & Graph Mastery — mixed-topic questions that don't tell you the pattern up front. Concept comparisons, when-to-use-what selection, and applied scenarios. The wrap-up that simulates real interview pressure.

The signals are well-defined once you know them. Reach for DP when the problem has overlapping subproblems (the same subcomputation appears across recursive calls) AND optimal substructure (an optimal solution can be built from optimal solutions to subproblems) AND the state space is small enough to enumerate. Reach for greedy when a local optimal choice provably leads to a global optimal (interval scheduling, fractional knapsack, MST), and verify the proof — most greedy mistakes come from trying it when the greedy property doesn't actually hold. Reach for brute force when the input size is small (n ≤ 20 for bitmask-able state, n ≤ 10 for permutation-based search). The Introduction to DP & Graph Algorithms topic in this course names the trigger conditions precisely.

State definition first, recurrence second, implementation third. The single highest-leverage skill in DP is defining the right state — what does dp[i] or dp[i][j] represent? Most DP failures aren't about coding; they're about defining a state that doesn't capture enough information to write the transition. The Abekus format helps because every DP MCQ in this course shows you the state definition and the recurrence explicitly in the explanation, so you internalise the pattern of "name the state, write the transition" across hundreds of problems.

Pair the MCQ practice with implementing 3–4 DP problems end-to-end in your language of choice every week. Not 30 — three to four, done carefully, with both memoisation and tabulation forms, plus space optimisation where applicable. After 6 weeks of this rhythm alongside the MCQs, you'll have the DP vocabulary that takes most self-taught learners a year.

You see one question at a time. It's usually a problem statement plus 4 candidate state definitions or recurrences, or a graph snippet with a complexity question. You pick the right state, the right recurrence, or the right algorithm. If you're right, you move on. If you're wrong, you get a written explanation — what the correct state was, why the alternative was wrong, what the complexity actually is. The platform tracks your accuracy per topic so you know where the gaps are. An AI guide is available alongside if you want a recurrence explained step by step.

The honest math: 1,554 MCQs at roughly 40 seconds per question works out to about 1,035 minutes — call it 17 hours. That's actual focused time. Most learners spread it across 3 to 5 weeks at 60–90 minutes per day. The DP-foundations layer (Introduction, DP Fundamentals, 1D DP) takes about 4 hours; the DP-patterns layer (2D DP, Knapsack, Sequence DP, Interval/Tree/Bitmask DP) takes 7–8 hours, with Sequence DP and Bitmask DP being the densest topics; the graph-algorithms layer (Shortest Paths, MST, Topological Sort & Union-Find, Advanced Graph, Mastery) closes in another 6 hours.

This is the slowest course in the DSA series to absorb. Plan 5–6 weeks if you're new to DP, even with Parts 1 and 2 finished. Most learners need to revisit the DP Fundamentals topic two or three times before state-definition becomes intuitive — that's normal and expected. Either timeline is fine — what matters is finishing all 12 topics.

This is Part 3 of Abekus's four-course DSA Mastery series. The natural sequence is: Part 1: Data Structures Fundamentals → Part 2: Algorithms & Core Patterns → this course → DSA — Interview Mastery (Part 4, mixed mock-style problems that simulate real coding rounds).

For language pairing, take this course alongside Python Fundamentals if your Python basics need work — most modern interview prep happens in Python and Abekus's question explanations assume language familiarity. The completion certificate from this course is publicly verifiable and can be added to your resume or LinkedIn profile.